Effects of Adverse Birth Events on ... -...
105
Effects of Adverse Birth Events on Maternal Mood, Maternal Functional Status and Infant Care by Diane F. Hunker Bachelor of Science in Nursing, University of Pittsburgh, 1989 Masters in Business Administration, University of Pittsburgh, 1994 Submitted to the Graduate Faculty of The School of Nursing in partial fulfillment of the requirements for the degree of Ph.D. in Nursing University of Pittsburgh 2007
Transcript of Effects of Adverse Birth Events on ... -...
Effects of Adverse Birth Events on Maternal Mood, Maternal Functional Status
and Infant Care
by
Diane F. Hunker
Bachelor of Science in Nursing, University of Pittsburgh, 1989
Masters in Business Administration, University of Pittsburgh, 1994
Submitted to the Graduate Faculty of
The School of Nursing in partial fulfillment
of the requirements for the degree of
Ph.D. in Nursing
University of Pittsburgh
2007
UNIVERSITY OF PITTSBURGH
School of Nursing
This dissertation was presented
by
Diane F. Hunker, BSN, MBA, RN
It was defended on
October 15, 2007
and approved by
Susan A. Albrecht, Ph.D., RN, FAAN
Ellen Olshansky, DNSc, RNC, FAAN
Katherine L. Wisner, M.D., M.S.
Thelma E. Patrick, Ph.D., RN
ii
Copyright © by Diane F. Hunker
2007
iii
Effects of Adverse Birth Events on Maternal Mood, Maternal Functional Status and
Infant Care
Diane F. Hunker, PhD, MBA, RN
University of Pittsburgh, 2007
Unplanned, adverse events during labor or delivery may generate a negative response for the
mother during the early postpartum period, resulting in disruption of usual functioning and
mood. Alterations in maternal mood can lead to a more debilitating condition known as
Postpartum Depression. Postpartum Depression negatively affects the quality of life and
functional status of mothers and infants. High levels of maternal depressive symptoms are
associated with parenting, infant attachment, behavioral problems and cognition (Beck 2002).
Little research has been completed exploring the relationship of adverse, unplanned events in
labor or delivery and maternal mood, functional status and infant care in the immediate
postpartum period. The purpose of this study was to examine the relationship of adverse events
in labor or delivery and mood, functional status and infant care at 2-weeks postpartum. The
secondary aim was to explore the role of social support as a possible moderator in the
relationship between adverse birth events and maternal outcomes. A secondary analysis of data
was performed using data collected in a descriptive, longitudinal study examining the effects of
antidepressant use during pregnancy. Participants included a convenience sample of 123
women. The main outcome measures included maternal mood, functional status, and infant care
at 2-weeks postpartum. Adverse events in labor or delivery did not significantly predict mood
(odds ratio =1.34, p=.536), functional status (R2 change = .001, p=.66), or infant care (R2
change=.004, p=.48) at 2-weeks postpartum when controlling for depression during pregnancy,
iv
antidepressant use at delivery, education level, age, and parity. Social support had significant
effects on mood (p=.02), functional status (p=.014), and infant care (p < .001) but did not
moderate the effect of adverse events when predicting mood (odds ratio=1.01, p=.045),
functional status (R2 change =.009, p=.056) and infant care (R2 change<.001, p=.92). The
occurrence of an adverse event in labor or delivery does not appear to predict alterations in
mood, functional status, or infant care at 2-weeks postpartum. Although social support does
appear to be related to mood, functional status and infant care, it does not appear to moderate the
effect of adverse events on the selected outcomes.
v
TABLE OF CONTENTS
PREFACE....................................................................................................................................XI
1.0 INTRODUCTION........................................................................................................ 1
1.1 PURPOSE............................................................................................................. 1
1.2 CONCEPTUAL FRAMEWORK....................................................................... 3
1.3 SPECIFIC AIMS ................................................................................................. 5
1.4 RESEARCH QUESTIONS................................................................................. 6
1.5 SIGNIFICANCE OF THE STUDY ................................................................... 6
2.0 BACKGROUND .......................................................................................................... 8
2.1 MATERNAL CHARACTERISTICS AND ADVERSE BIRTH EVENTS ... 8
2.2 SOCIAL SUPPORT .......................................................................................... 20
2.3 MATERNAL MOOD ........................................................................................ 22
2.4 FUNCTIONAL STATUS.................................................................................. 28
2.5 INFANT CARE.................................................................................................. 29
2.6 SIGNIFICANCE OF THE EARLY POSTPARTUM PERIOD ................... 30
2.7 SUMMARY........................................................................................................ 31
3.0 METHODS ................................................................................................................. 33
3.1 DESIGN.............................................................................................................. 33
3.2 SAMPLE............................................................................................................. 34
vi
3.3 VARIABLES...................................................................................................... 36
3.4 MEASURES ....................................................................................................... 36
3.5 DATA MANAGEMENT................................................................................... 42
3.6 ANALYSIS ......................................................................................................... 42
4.0 PILOT STUDY........................................................................................................... 47
4.1 RELIABILITY AND VALDITY ESTIMATES FOR THE EDINBURGH
POSTNATAL DEPRESSION SCALE USED AT 2 WEEKS POSTPARTUM........... 47
4.1.1 PURPOSE....................................................................................................... 47
4.1.2 METHODS..................................................................................................... 48
4.1.3 RESULTS....................................................................................................... 53
4.1.4 DISCUSSION................................................................................................. 56
5.0 RESULTS ................................................................................................................... 61
5.1 SAMPLE CHARACTERISTICS..................................................................... 61
5.2 DESCRIPTIVE STATISTICS ......................................................................... 63
5.2.1 INDEPENDENT VARIABLE: ADVERSE EVENTS................................ 63
5.2.2 DEPENDENT AND MODERATING VARIABLES: MATERNAL
MOOD, FUNCTIONAL STATUS, INFANT CARE, SOCIAL SUPPORT......... 65
5.3 PRIMARY AIM................................................................................................. 66
5.3.1 MATERNAL MOOD.................................................................................... 66
5.3.2 FUNCTIONAL STATUS.............................................................................. 68
5.3.3 INFANT CARE.............................................................................................. 69
5.4 SECONDARY AIM........................................................................................... 70
5.4.1 MAIN EFFECT OF SOCIAL SUPPORT................................................... 71
vii
5.4.2 MODERATING EFFECT OF SOCIAL SUPPORT ................................. 71
6.0 DISCUSSION ............................................................................................................. 75
6.1 CHARACTERISTICS OF THE SAMPLE..................................................... 75
6.2 DISCUSSION OF RESULTS ........................................................................... 76
6.2.1 ADVERSE BIRTH EVENTS ....................................................................... 77
6.2.2 FUNCTIONAL STATUS.............................................................................. 78
6.2.3 MATERNAL MOOD.................................................................................... 79
6.2.4 SOCIAL SUPPORT ...................................................................................... 80
6.3 CLINICAL SIGNIFICANCE........................................................................... 80
6.4 LIMITATIONS.................................................................................................. 81
6.4.1 CHOSEN OUTCOMES................................................................................ 81
6.4.2 STUDY DESIGN ........................................................................................... 81
6.4.3 SAMPLE......................................................................................................... 82
6.5 FUTURE STUDIES........................................................................................... 82
6.5.1 TIMING.......................................................................................................... 82
6.5.2 INSTRUMENTS............................................................................................ 83
6.5.3 DEFINITION OF ADVERSE EVENTS ..................................................... 84
6.5.4 CONCEPT MEASUREMENT..................................................................... 84
6.6 CONCLUSION .................................................................................................. 85
APPENDIX. IRB APPROVAL.................................................................................................. 86
BIBLIOGRAPHY....................................................................................................................... 88
viii
LIST OF TABLES
Table 1. Adverse, Unplanned Events Identified from the Peripartum Events Scale..................... 39
Table 2. Participant Demographics for Pilot Study (Total n=229) ............................................. 49
Table 3. Frequency Data for Pilot Study ...................................................................................... 54
Table 4. EPDS Reliability Data .................................................................................................... 55
Table 5. EPDS Validity Data ........................................................................................................ 56
Table 6. Spearman’s Rho Inter-Item Correlation for All Scores.................................................. 58
Table 7. Demographic Characteristics of Sample (n=123) ......................................................... 62
Table 8. Adverse, Unplanned Events Identified in the Sample ..................................................... 64
Table 9. Measures of Central Tendency for the Independent Variables ...................................... 65
Table 10. Logistic Regression Results for Maternal Mood .......................................................... 67
Table 11. Multiple Linear Regression Coefficients for Functional Status ................................... 69
Table 12. Multiple Linear Regression Coefficients for Infant Care ............................................ 70
Table 13. Logistic Regression Results for Maternal Mood and Social Support........................... 72
Table 14. Multiple Linear Regression Coefficients for Functional Status and Social Support.... 73
Table 15. Multiple Linear Regression Coefficients for Infant Care and Social Support ............. 74
ix
LIST OF FIGURES
Figure 1. Conceptual Framework for the Study of Adverse Birth Events, Maternal Mood,
Maternal Functional Status, and Infant Care................................................................................. 1
x
PREFACE
I would like to begin by thanking my husband, David Hunker, for supporting my idea and
joining in with the rest of my family on undertaking many sacrifices over the past several years.
There are many faculty members at the University of Pittsburgh who have made this degree
possible. I would first like to thank Dr. Susan Albrecht for taking time out of her busy July day
to meet with me and convince me to go back to graduate school –not for a Masters, but for a
Ph.D. She then went on to be an integral part of my dissertation committee and my future career
search. I would like to thank the graduate faculty at the School of Nursing, including Dr. Ellen
Olshansky, who graciously agreed to be a part of my dissertation committee. Each and every
professor had something special to offer during the pursuit of my doctoral education. I would
like to thank. Dr. Katherine Wisner, and her team. Without their support, I would not have been
able to gain the substantial amount of knowledge, and complete my research in the amount of
time that I did. And finally, from the University of Pittsburgh and UPMC Magee-Womens
Hospital, I would like to extend my heartfelt gratitude to Dr. Thelma Patrick. Once I got started
in the Ph.D. program, Dr. Patrick provided consistent and meaningful support and mentoring. I
also attribute my success and extend much thanks to the many family members and colleagues
who supported my endeavors. And finally, I would like to dedicate this dissertation to my sons,
xi
Dylan, Drew, and Joseph. They continue to be the inspiration for everything I do. I hope that by
completing this program, they too will be able to see that goals and perseverance do pay off.
xii
1.0 INTRODUCTION
1.1 PURPOSE
The purpose of this study was to examine the relationship of adverse birth events on maternal
outcomes in the early postpartum period. The media often portrays a very idyllic view of the
birthing event, with joyful pregnancies, seemingly effortless labor, and uneventful, quick
deliveries. However, data regarding birth events are not consistent with this portrayal.
Unexpected complications occur in labor or deliveries that are often confusing to the pregnant,
laboring woman. In 2005, over 210,000 deliveries were the result of precipitous, prolonged or
dysfunctional labor (National Vital Statistics Report, 2006). Breech or atypical fetal
presentations accounted for over 170, 000 births, and fetal distress in labor was evident in over
132,000 deliveries in 2005 (National Vital Statistics Report, 2006). Eight percent of all deliveries
are diagnosed with dystocia in some form (Algovik, et al, 2004).
paragraph.
Vaginal delivery is the expected mode of delivery, and women are often ill-prepared for
circumstances that result in more aggressive intervention. In childbirth classes, women are told
that one in four women is likely to have a cesarean delivery (Fawcett, et al, 2005). In spite of
this incidence, qualitative analysis of post delivery interviews confirmed that women felt
1
unprepared for operative delivery and thought that their birth plan or antenatal classes had not
prepared them adequately for the birth event (Murphy, et al, 2003).
As noted by Waldenstrom, a woman's dissatisfaction with her birth may affect her
emotional well-being and willingness to have another baby (Waldenstrom, et al, 2004).
Alterations in maternal mood can go largely unrecognized during the postpartum period. Most
people consider the “baby blues” a normal part of childbearing; however, alterations in maternal
mood can lead to a more debilitating condition known as Postpartum Depression (PPD).
Prevalence rates of PPD are estimated at 14.5% (Gavin, et al). In other words,
approximately one out of seven new mothers experiences depression (Gaynes, et al, 2005). PPD
affects 500,000 mothers and infants each year with subsequent recurrence rates in future
pregnancies at about 25% ((Wisner, et al, 2002; Wisner, et al, 2004).
Altered maternal mood can have negative effects for mothers including reduced quality
of life, reduced functional status, and altered parenting skills (Beck, 2002). If left untreated,
altered maternal mood can result in prolonged psychiatric illness and more serious detriment for
the entire family. Altered maternal mood can also have negative effects for infants including
reduced mother-infant attachment, increased infant behavioral problems, and decreased infant
cognition (Beck, 2002). Early screening of altered maternal mood, and treatment of PPD, is
critical in reducing potential downstream effects on infant development and future psychiatric
illness in children.
This study focused on a variety of negative birth events in contrast to prior studies that
have specified either cesarean section or assisted vaginal delivery compared to spontaneous
vaginal delivery. The concepts central to this exploration are depicted in the conceptual
framework for the study (Figure 1).
2
1.2 CONCEPTUAL FRAMEWORK
An adverse birth event is likely to precipitate alterations in maternal mood, functional status, and
infant care. There is a need to focus on alterations in maternal mood early in the postpartum
period beyond what is typically expected. Alterations in maternal mood include depressive
symptoms which can include fatigue (beyond what is expected for a new mother), feelings of
worthlessness or guilt, diminished ability to concentrate, insomnia, or diminished interest or
pleasure (Wisner, et al, 2004). Alterations in functional status after childbirth include the
inability to perform tasks at the level of functioning that was exhibited prior to childbirth. Such
Maternal
Infant Care
Maternal Functional Status
Adverse
Age, Parity, Education Level, Hx of depression in
Pre
Social Support
gnancy,
Figure 1. Conceptual Framework for the Study of Adverse Birth Events, Maternal Mood, Maternal Functional Status, and Infant Care
3
activities may include household chores, social and community activities, self-care activities and
occupational activities (Fawcett, et al, 1998). Infant care, although not exhibited prior to
childbirth, is a primary goal of successful functional status after childbirth. It is possible that
women who had negative birth experiences may need to meet their own emotional needs before
they can meet those of their infants, or may feel obligated to meet the infant’s needs, but not their
own (Kennell, et al ,1979).
Previous studies have shown that social support is a moderating factor in postpartum
recovery. Social support, defined as a well-intentioned action that is given willingly to a person
with whom there is a personal relationship and that produces an immediate or delayed positive
response in the recipient, has been studied extensively in pregnancy and parenting (Hupcey,
1998). Receiving social support early in the postpartum period has been linked to better
maternal-child interactions and smaller incidence of PPD (Logsdon, et al, 1994). Prior research
findings suggest several intervention points for new mothers including adequate social support
screening (Soet, et al, 2003).
The acknowledgement that other factors can also greatly impact a mother’s birth
outcomes were addressed as covariates influencing the maternal response. Maternal age, parity
and education level can influence a mother’s reaction to her birth. Also, prior history of
depression and antidepressant use during delivery were also examined as they could directly
contribute to a mother’s postpartum mental state.
By taking into account covariates which are thought to influence the maternal outcomes,
identifying actual adverse birth events from the medical record, and measuring social support
which is believed to moderate the maternal outcomes, this study was able to determine if there is
in fact a relationship between unplanned, adverse birth events and specific maternal outcomes.
4
In this study, both women who have experienced an adverse birth event, as well as women with
uncomplicated vaginal deliveries, were investigated in an attempt to determine if there is a
significant relationship with maternal mood, functional status and infant care in the early
postpartum period.
1.3 SPECIFIC AIMS
The purpose of this study was to examine the relationship of adverse birth events on maternal
outcomes in the early postpartum period. More specifically, the primary aim was to examine the
relationship between unplanned, adverse birth events during labor or delivery and maternal
mood, maternal functional status, and infant care at 2 weeks postpartum. It was hypothesized
that unplanned, adverse birth events were associated with altered maternal mood, decreased
maternal functional status and decreased comfort in infant care at 2 weeks postpartum.
The secondary aim for this study was to explore if social support was a moderator in the
relationship between unplanned, adverse birth events and maternal mood, functional status, and
infant care at 2 weeks postpartum. Previous studies have shown that social support is a
moderating factor in postpartum recovery. It was hypothesized that social support influenced the
maternal response at 2 weeks postpartum to negative birth events.
5
1.4 RESEARCH QUESTIONS
The research questions for this study were as follows:
1. Is there a relationship between unplanned, adverse birth events and maternal
mood, maternal functional status, and infant care at 2 weeks postpartum?
2. Does social support serve as a moderator in the relationship
between unplanned, adverse birth events and maternal mood,
maternal functional status, and infant care at 2 weeks
postpartum?
1.5 SIGNIFICANCE OF THE STUDY
Little research has been completed exploring the relationship of actual adverse, unplanned events
in labor or delivery and maternal outcomes, specifically maternal mood, functional status and
infant care, in the immediate postpartum period. Much of the literature addresses high-risk
pregnancy events, such as preterm labor, that have obvious long-term consequences. However,
when the pregnancy has complications, the events in labor or delivery are often expected. When
the pregnancy is uncomplicated, adverse events in labor or delivery are often not expected.
6
Understanding the relationship between unplanned, adverse birth events and maternal
outcomes should assist providers in making decisions regarding treatment during the immediate
postpartum period. By identifying and clarifying the need for such treatments during the
immediate postpartum period, healthcare providers can seek to minimize the negative effects of
traumatic births and potentially decrease long- term morbidity among postpartum women. In
addition to improving patient outcomes, health care costs can be minimized by eliminating the
need for further care and evaluation.
7
2.0 BACKGROUND
2.1 MATERNAL CHARACTERISTICS AND ADVERSE BIRTH EVENTS
The purpose of this study was to examine the relationship between unplanned, adverse
birth events and maternal outcomes. Adverse birth events in labor or delivery may result in a
negative birth experience. Negative birth experiences have been the subject of both quantitative
and qualitative research in the past. The birth of a child is a life-altering and memorable event.
In fact, women can accurately remember details of their first births even 20 years later (Simpkin,
1992). Not all births are positive experiences. In a prospective, longitudinal study done by
Creedy, et al, examining women (n=499) in the third trimester and following them until 4 to 6
weeks postpartum, it was discovered that one in three women (33%) identified a negative birth
event and reported the presence of at least three trauma symptoms following the birth of their
infant (Creedy, et al, 2000). When research studies have considered the question of negative
birth experiences, the general paradigm is to compare emotional reactions to cesarean sections
and vaginal births; however, the types and circumstances surrounding a negative birth experience
can encompass many different events. In 2004, Waldenstrom completed a longitudinal study
investigating the prevalence and risk factors of a negative birth experience in a national sample
(Waldenstrom, et al, 2004). In this study, it was found that risk factors for a negative birth
experience were related to unexpected medical problems and participants' social background.
8
There are many examples of unplanned, adverse birth events that can occur during labor
or delivery. A prime example of an unplanned, adverse birth event in labor or delivery includes
the use of forceps. Forceps are an instrument varying in size and shape but typically
encompassing 2 crossing branches, or blades, that are designed for extraction of the fetus while
in vertex position (Cunningham, et al, 1993). The practice of using forceps for assisted delivery
is losing popularity as new OB/GYNs are not being adequately trained in their use, and vacuum
extraction has become the method of choice for assisted vaginal delivery.
A vacuum extractor, also used to assist vaginal delivery, is a device that attaches traction
by suction to the fetal scalp. The fetus is then manually extracted from the birth canal by the
physician (Cunningham, et al, 1993). Use of forceps to assist with vaginal delivery, as well as the
use of a vacuum extractor, is an unplanned, often painful event that may leave the mother in need
of significant physical attention following her delivery. Often these procedures are related to
prolonged second stage of labor, birth weight and head circumference, and can result in
extensive perineal lacerations involving the anal sphincter (Hudelist, et al, 2005). Usually
combined with medio-lateral episiotomies, these types of deliveries can result not only in trauma
to the mother, but can have grave consequences to the infant’s physical well-being as well
(Hudelist, et al, 2005). Thus, leaving the mother not only in distress regarding her own
emotional and physical trauma, but also in distress because of her infant’s status following
delivery.
Another type of unplanned, adverse event in labor or delivery is the need for an
unplanned caesarean section. Typical reasons cited for an unplanned cesarean section during
labor are as follows:
• Failure to progress (FTP) in labor (insufficient cervical dilatation and effacement);
9
• Cephalo-pelvic disproportion (CPD) meaning the fetal head does not fit in through the
mother’s pelvis;
• Failure to descend (FTD) into the birth canal (inadequate station). (Cunningham, et al,
1993).
A prolonged stage of labor may be also interpreted as an unplanned, adverse event. There
are three stages of labor. The first stage is defined as the commencement of uterine contractions
having sufficient frequency, intensity and duration to bring about cervical changes and ends
when the cervix is fully dilated (Cunningham, et al, 1993). The second stage begins when
dilatation is complete and ends with delivery of the fetus. During this stage, the mother is usually
‘pushing’. The third stage begins immediately after the delivery of the fetus and ends with the
delivery of the placenta and fetal membranes (Cunningham, et al, 1993).
The average duration of the first stage of labor is about 8 hours for nulliparous women
and 5 hours for multiparous women (Cunningham, et al, 1993). In this study, prolonged first
stage of labor was defined as exceeding 20 hours for nulliparous women and exceeding 14 hours
in multiparous women. The median duration of the second stage of labor in nulliparous women
was 50 minutes and 20 minutes for multiparous women (Cunningham, et al, 1993). Prolonged
second stage of labor was defined as exceeding 2 hours in nulliparous women and 1.5 hours in
multiparous women.
As opposed to prolonged labor, women can also suffer negative effects from precipitate
labor. Precipitate delivery is defined as extremely rapid labor and delivery and usually is the
result of abnormally low resistance of the soft parts of the birth canal accompanied by
abnormally strong contractions (Cunningham, et al, 1993). Occasionally, precipitate delivery is
the result of absence of painful sensations and thus a lack of awareness of vigorous labor
10
(Cunningham, et al, 1993). In either case, the presence of an intense, unexpected labor process
can be considered an adverse labor event. For this study, precipitate delivery was defined as
labor and delivery occurring in less than 3 hours.
Significant blood loss or extensive perineal lacerations are other types of unplanned,
adverse events. Significant blood loss after the third stage of labor has been historically defined
as 500 cubic centimeters or more, although this figure is probably low considering clinicians do
not have accurate ability to quantify blood loss (Cunningham, et al, 1993).
The concept of significant perineal lacerations can be reserved for third and fourth degree
vaginal lacerations. These more serious types of lacerations are often associated with an
episiotomy, and less significant (first and second degree) lacerations (Cunningham, et al, 1993).
Third degree lacerations extend through the skin, mucous membrane and perineal body, and
involve the anal sphincter. Fourth degree lacerations are distinguished by a third degree tear that
extends through the rectal mucosa to expose the lumen of the rectum (Cunningham, et al, 1993).
Tears of the urethra are also likely to occur with fourth degree lacerations and may bleed
profusely. Repair of third and fourth degree lacerations is complicated by the irregular lines of
tissue cleavage. Issues with healing and post-operative pain are often more significant with these
types of lacerations (Cunningham, et al, 1993).
Placental abruption is another example of an unplanned, adverse event. Placental
abruption is defined as the separation of the placenta from its site of implantation before the
delivery of the fetus. Bleeding from placental abruption lies between the placental membranes
and the uterus and then escapes through the cervix causing an external hemorrhage
(Cunningham, et al, 1993). Occasionally, the bleeding remains internal between the placental
11
membranes and the uterus, causing an internal hemorrhage and more serious consequences
(Cunningham, et al, 1993).
The presence of fetal distress as evidenced by fetal heart rate monitoring is another type
of adverse event in labor. Often very subjective to interpretation and difficult to diagnose, fetal
distress is a diagnosis based on fetal heart rate patterns. Often terms ‘reassuring’ and ‘non-
reassuring’ heart rate are used instead to eliminate some of the uncertainty in the diagnosis
(Cunningham, et al, 1993). Fetal heart rate patterns are dynamic and can change rapidly
(Cunningham, et al, 1993). The patterns are usually termed fetal distress when the physician can
no longer assuage doubts about the condition of the fetus, and is no longer confident that the
fetus is not being compromised (Cunningham, et al, 1993). A variety of interventions can occur
in labor once it is believed fetal well-being is being compromised, and emergency cesarean
section or assisted vaginal delivery may result (Cunningham, et al, 1993).
Another type of adverse event can involve issues with the amniotic fluid in labor.
Amniofusion is a therapy designed to improve symptoms of fetal distress by infusing warm
saline via an intrauterine pressure catheter into the uterus (Cunningham, et al, 1993). Fetal
distress in this case could have been caused by meconium-staining, compressed cord, or
oligohydramnios – a decreased amount of amniotic fluid. Amniofusion has been gaining in
popularity as the rates of fetal scalp blood sampling are steadily declining. Meconium-staining is
the presence of fetal meconium in the amniotic fluid. Occurrence rates of meconium-staining are
about 12 -20 percent of all pregnancies (Cunningham, et al, 1993). Neonatal morbidity and
mortality can be associated with meconium-staining as a result of fetal aspiration (Cunningham,
et al, 1993). Requirement of an amniofusion in labor or meconium-staining of the amniotic fluid
can be considered adverse, unplanned events that occur in labor.
12
Preeclampsia that first presents in labor is another example of an adverse event in labor.
Pregnancy-induced hypertension can be categorized as preeclampsia when there is the
development of hypertension plus proteinuria, and/ or edema in the mother (Cunningham, et al,
1993). In some cases, signs of preeclampsia do not present until the women is in labor and
subsequently require aggressive intervention, and possibly emergent delivery to alleviate the
symptoms and decrease potentially dangerous maternal and neonatal effects.
Another example of an adverse event is the occurrence of a cord prolapse in labor.
Umbilical cord prolapse is a serious complication that is facilitated by an imperfect adaptation
between the presenting fetal part and the mother’s pelvic inlet causing the cord to prolapse
through the birth canal (Cunningham, et al, 1993). Sometimes it is a result of artificial rupture of
the amniotic membranes in labor, or a fetal breech presentation (Cunningham, et al, 1993).
Failure to deliver the fetus quickly can result in fetal demise.
The final example of an unplanned, adverse event is a shoulder dystocia during the
second stage of labor. Shoulder dystocia occurs in less than 2% of all deliveries and severity
varies by definition (Cunningham, et al, 1993). Caused by macrosomia, or an increase in fetal
body size in relation to the fetal head size, it occurs when the head delivers, but the body remains
inside the birth canal. There are various procedures to remedy the crisis. The occurrence of a
shoulder dystocia often results in significant fetal morbidity and mortality, especially if not
intervened correctly (Cunningham, et al, 1993).
Although I have mentioned the most common types of unplanned events in labor, there
are many, more obscure events that can occur. Most adverse events are not mutually exclusive,
thus a woman may have more than one event. For instance, prolonged second stage of labor
13
could result in the need for a vacuum extraction. Another example could be preeclampsia that
first presents in labor and necessitates the need for an unplanned, cesarean section.
Term infants are often required to go to the neonatal intensive care unit (NICU) following
an unplanned, adverse birth event. Most women probably believe that if all of their prenatal
screening procedures were unremarkable, and they made it past 37 weeks gestation, their infants
will be fine. The emotional effects of a term infant needing to get specialized care following a
delivery event can be detrimental to a mother’s well-being and can impact her ability and
willingness to adequately care for her needs (Waldenstrom, et al, 2004).
The consequences of an adverse birth event can be evident across the continuum of
physical to mental health. The progress of recognizing the impact of adverse birth events has led
to explorations of the physical effects and method of delivery. Lydon-Rochelle, Holt and Martin
assessed the association between method of delivery and the general health status, sexual, bowel
and urinary functioning of primiparous women as measured at 7 weeks postpartum (Lydon-
Rochelle, et al, 2001). At 7 weeks postpartum, women who had cesarean or assisted vaginal
deliveries reported significantly lower postpartum general health status scores than women with
unassisted vaginal delivery (e.g., did not use forceps or vacuum extraction). Additionally,
women with assisted vaginal deliveries reported significantly worse sexual, bowel and urinary
functioning. The results suggest that more careful attention to the postpartum general health and
sexual functioning of women with cesarean and assisted vaginal deliveries may be merited.
Similar conclusions regarding method of delivery were found in subsequent studies.
Murphy, Pope, Frost, and Liebling completed a qualitative study (n=27 women) and determined
that operative delivery had a noticeable impact on women's views about future pregnancy and
delivery (Murphy, et al, 2003). Three important conclusions were reached by the investigators of
14
this study: (1) antenatal variables did not contribute to the development of acute or chronic
trauma symptoms, (2) women who experienced both a high level of obstetric intervention and
dissatisfaction with their intrapartum care were more likely to develop trauma symptoms than
women who received a high level of obstetric intervention or women who perceived their care to
be inadequate, and (3) Posttraumatic Stress Disorder (PTSD) is evident in women after a
childbirth experience that include intrusive obstetric intervention during labor and delivery, and
the care provided to women in labor.
Continuing with research in the area of method of delivery and physical effects, Hudelist,
et al, examined risk factors for third and fourth degree perineal tears in patients having either a
spontaneous vaginal delivery or an assisted vaginal delivery with forceps combined with a
medio-lateral episiotomy (Hudelist, et al, 2005). Their retrospective, case-control study of 5,377
women revealed that nulliparity, prolonged first and second stage of labor, high birth weight, and
type of delivery were identified as risk factors for perineal trauma. When the risk factors were
entered into a multivariate regression model, high birth weight and episiotomy in conjunction
with forceps proved to be the most significant risk factors for trauma. Perineal trauma sustained
as a result of unplanned events, such as assisted vaginal birth, proved to have a significant effect
on women in the postpartum period; thus, potentially interfering with their postpartum functional
status.
Similarly, Lydon-Rochellle, et al, studied the association of method of delivery, including
unplanned cesarean section and assisted vaginal delivery, and the rate of maternal
rehospitalization in a retrospective cohort study (n=256,795) (Lydon-Rochelle, et al, 2000).
Compared with women who had spontaneous vaginal deliveries, women with cesarean sections
and assisted vaginal delivery were significantly more likely to be rehospitalized. It can be
15
inferred from their findings that rehospitalization could have potential negative effects on both
the mother’s functional status and general mental well-being.
Research has shown that women who had emergency cesarean sections in labor following
complete cervical dilatation were more likely to have intra-operative trauma and neonatal
asphyxia (Allen, et al, 2005). Although there is a body of literature regarding outcomes with
elective cesarean sections, Allen’s, et al, study supports the suggestion that emergency cesarean
delivery, particularly when complete cervical dilatation has occurred, has significant effects on
maternal morbidity.
Gardella, et al, also completed a retrospective, cohort study examining the effects of
sequential use of vacuum and forceps for assisted vaginal delivery on both neonatal and maternal
outcomes (Gardella, et al, 2001). Their study compared 3,741 women who had both vacuum and
forceps assistance in their deliveries with 11, 223 women with spontaneous vaginal deliveries.
They found that the risk of using both instruments greatly increased the risk of neonatal and
maternal outcomes, such as neonatal intracranial hemorrhage, nenonatal brachial plexus injury,
neonatal facial nerve damage, maternal perineal trauma, hematoma and postpartum hemorrhage.
The findings from their study suggest a synergistic interaction effect of multiple instrument use
during a single delivery; thus, reinforcing the premise that multiple unplanned events in delivery
could be even more detrimental than a single unplanned event.
Similar to physical effects as a result of labor and delivery events, many researchers have
also studied mental health effects of unplanned, adverse events. Koo, et al, completed a
retrospective cohort study examining the risk of alterations in maternal mood after emergency
delivery (Koo, et al, 2003). Using the Edinburgh Postnatal Depression Scale (EPDS) as their
measure and chi square statistical analysis between groups, they found that at 6 weeks
16
postpartum, when compared with women having non-emergency deliveries (n=191), women
having emergency delivery (n=55) had about twice the risk of developing depressive symptoms.
The main finding of their study suggested that women who underwent an emergency delivery
have an increased risk for developing PPD and that postnatal follow up of these women could be
beneficial in screening for alterations in mood.
In contrast, Patel, et al, found that in their prospective cohort study (n=14,663), women
who planned a spontaneous vaginal delivery and required an emergency delivery, whether it be a
cesarean section or assisted vaginal delivery, were not significantly at increased risk for postnatal
depression at 8 weeks postpartum when compared to women with spontaneous vaginal deliveries
(Patel, et al, 2005). Similarly, they found that elective cesarean section did not increase the risk
for PPD; thus, emphasizing the importance of women’s ability to make informed decisions
regarding the association of mode of delivery and PPD.
Opposed to studying maternal outcomes following mode of delivery, van de Pol, et al,
studied psychosocial factors as predictors of mode of delivery on a group of 354 healthy
nulliparous women (van de Pol, et al, 2006). They found that the only significant predictors of
operative delivery after spontaneous onset of labor are high birth weight, non-occiput anterior
presentation, advanced gestational age, and fetal distress during labor. Their findings confirm
that certain psychosocial traits, such as level of social support, depressive symptoms in
pregnancy, and specific personality traits in pregnancy, in pregnancy although they can affect
postpartum outcomes, do not predict or protect against assisted vaginal delivery or emergency
cesarean section.
Similarly, Wu, et al, studied whether or not mood disturbances in pregnancy were
associated with higher frequency of cesarean or assisted vaginal delivery (Wu, et al, 2002). Of
17
their study of 264 women, there was no statistically significant difference between the rate of
cesarean section or assisted vaginal delivery in women with and without elevated depression
scores in pregnancy.
Womens’ perceptions, rather than the actual medical acknowledgement of an adverse
event, may impact a women’s response to the event. For example, if a new mother has a
prolonged second stage of labor that results in a vacuum extraction, her labor would be
considered as having an adverse event as defined in this study. However, for the first time
mother, this may appear normal and part of the labor and delivery process. In this study, we are
not considering the mother’s perception of her labor and delivery, but rather just the presence of
one or more adverse, unplanned events as recorded on the woman’s medical record.
Beck (2004), following her phenomenologic study investigating women’s experiences of
birth trauma, concluded that birth trauma “lies in the eyes of the beholder” meaning that mothers
often viewed their birth as traumatic, whereas physicians viewed the birth as routine, and vice-
versa. Such findings, although limited in scope, may warrant future studies including the
mother’s perception of her labor and delivery to determine if there is a relationship between the
mother’s perception and the actual events experienced, as well as to determine if there is a
relationship between the mother’s perception and the specified maternal outcomes.
Similarly, there may be psychosocial or physical health circumstances that influence the
maternal response to an adverse birth event. For instance, infertility may be one of those
circumstances in which the feelings of elation in finally achieving a pregnancy that results in the
delivery of the highly desired infant may surpass the definition of any aspect of the birth as
adverse. In this study, evaluation of certain circumstances, such as infertility, was not being
18
considered; therefore potentially minimizing, or in some cases exaggerating, the effect of the
adverse birth event.
Given the available research evidence, the following factors merit acknowledgement as
significant risk factors for alterations in maternal outcomes, particularly mood in the early
postpartum period: presence of prenatal depression symptoms which may or may not require
antidepressant use up until delivery, inadequate social support, very young mothers, and lower
socio-economic status (Horowitz & Goodman, 2004). As a result, those factors have been
selected to serve as covariates influencing the maternal outcomes identified for this study.
Research regarding history of depression both pre-pregnancy and prenatally will be
discussed further in the next section. Both factors have been the strongest predictors to date
regarding development of mood alterations in the postpartum period. Due to the large amounts of
literature supporting social support’s influential role in maternal outcomes, it seemed probable
that it would serve as a moderator in the relationship between birth events and maternal
outcomes. In addition, when considering the outcome of infant care particularly, parity seemed to
be a likely covariate. Having had a child previously, women should be more knowledgeable
regarding infant care, and perhaps more aware of their expected functional status and more
prepared to function in their required roles. Education level, the measure used in this study, has
often been representative of socio-economic status in research. Finally, maternal age was
selected as a covariate not only because of support from the literature noting differences with
younger mothers, but because of the assumption that women in their late teens may not have had
the same life experience preparing them for the life changes that are expected to occur with the
birth of a child. There are noted differences between maturity levels of women in their early 20s
19
as opposed to women in their mid-30s. Therefore, maternal age was selected as covariate
influencing the maternal mood, functional status and infant care in the early postpartum period.
2.2 SOCIAL SUPPORT
A significant body of research has demonstrated that social support is facilitative of well-being
during major life transitions and periods of acute stress (Cohen & Wills, 1985). Childbirth should
be considered a major life transition and as such it can be assumed that social support is
potentially influential in developing maternal outcomes following delivery. Historically, social
support, a multi-faceted and widely studied concept, has been difficult to define, conceptualize
and measure (Hupcey, 1998). There are many definitions for social support in the literature,
although many of them tend to be vague and simplistic. There are various models of social
support and many different approaches used for examining the concept (Hupcey, 1998). Social
support has been measured in a variety of ways including perceived support, adequacy of support
network, and satisfaction with support (Logsdon & Winders, 2003).
Sources of social support desired in childbirth have been studied. Although prior research
has found that the source of support differed based on age, socio-economic status and ethnicity,
the partner (baby’s father) was an important part of emotional, informational, and instrumental
support (McVeigh, 2000). Similarly, support given by health care providers and delivery
attendants has been shown to be influential.
Social support in labor and postpartum has been studied in conjunction with maternal
mood, functional status and infant care. Social support has also been studied with the broader
20
concept of mothering and has been found to cushion the experience of moving into the
motherhood role and critical for maternal role adaptation (McVeigh, 2000). In a prospective,
longitudinal study completed by McVeigh examining the relationship between satisfaction with
social support and functional status after childbirth, it was reported that satisfaction from support
from one’s partner was significantly correlated with infant care at 6 weeks postpartum, self-care
at 3 months, and community activities at 6 months (McVeigh, 2000). Similarly, Cooper, et al,
completed a descriptive, cross-sectional study on 147 South-African women and found that
social support provided by the partner was significantly increased in the non-depressed women
(Cooper, et al, 1999).
Previous studies have shown that social support is a major factor in postpartum recovery
and is strongly linked to alterations in maternal mood. Receiving social support early in the
postpartum period has been linked to smaller incidence of postpartum depression (Olshansky &
Sereika, 2005). A study completed by Coffman, et al, examined relationships in mothers of
distressed and normal newborns (Coffman, et al, 1999). The results of their study provided
support for a proposed model of social support indicating that close support and met expectations
of support were related to both maternal postpartum mood and positive maternal attitude toward
the infant. In fact, they found that maternal mood actually mediated the attitudes toward infants
thus reinforcing prior research suggesting that maternal mood contributes to disturbances in the
mother-infant relationship, and that postpartum social support has implications for both maternal
and child well-being.
Klaus’ continued program of research on mother-infant relationships supports the notion
that social support ante-and post partum improves the psychological health of the mother with
associated parental benefits which result in stronger mother-infant ties (Klaus, 1998).
21
2.3 MATERNAL MOOD
There have been many studies focused on alterations in maternal mood. For instance, alterations
in mood have been associated with immediate postpartum screening in an effort to wane long-
term detrimental effects. Recent focus in research in the area of maternal mood has been on
Posttraumatic Stress Disorder (PTSD) following an adverse birth event. Typically these studies
focused on later periods during the postpartum recovery where symptoms of PTSD, such as
nightmares and flashbacks, become more evident and more debilitating.
PTSD was first recognized as a distinct diagnosis based on a traumatic life event in the
Diagnostic and Statistical Manual –III (DSM-III) (Bailham & Joseph, 2003). Initially, PTSD was
defined as something outside the range of human experience, leading to the confusion by
clinicians regarding inclusion of childbirth to the definition. Recently childbirth has been
recognized as a possible event that can trigger PTSD, and has since been acknowledged as such
in the DSM-IV 1994 edition (Bailham & Joseph, 2003). Several unique features of PTSD
following childbirth include sexual avoidance, mother-infant attachment problems, difficulty
breastfeeding, difficulty bonding, and related parenting problem (Bailham & Joseph, 2003).
A number of risk factors have evolved from the literature for PTSD following difficult
childbirth including type of delivery. It has been suggested that emergency cesarean section is
likely to be associated with postpartum emotional difficulties (Bailham & Joseph, 2003).
Instrumental delivery may also be a significant risk factor for PTSD. Social support has been
found to be an important protective factor in PTSD-prone populations, and interventions
22
involving the provision of support appear to be effective in promoting mental health following
childbirth (Bailham and Joseph, 2003). Similarly, in a review of evidence for PTSD after birth,
Susan Ayers confirmed that the research suggests that a history of psychiatric problems, mode of
delivery, and low social support during labor put women at increased risk of postnatal PTSD
(Ayers, 2004)
Theoretical models of PTSD acknowledge the role of individual differences and the
influence of psychosocial factors which can explain why some women experience PTSD and
others do not. There is some noted overlap in symptoms between PTSD and PPD. However, it is
believed that these are distinct disorders and it is possible that women present with one without
the other.
Creedy, Shochet, and Horsfall studied psychological health following an adverse birth
event (Creedy, et al, 2000). Twenty-eight women (5.6%) met DSM-IV criteria for acute
Posttraumatic Stress Disorder. The level of obstetric intervention experienced during childbirth
(beta = 0.351, p < 0.0001) and the perception of inadequate intrapartum care (beta = 0.319, p <
0.0001) during labor were consistently associated with the development of acute trauma
symptoms exhibited with PTSD.
Continuing with the focus on PTSD, Cheryl Beck used descriptive phenomenology to
explore the essence of mother’s experiences of PTSD after childbirth (Beck, 2004). This study
examining the lives of mothers with Posttraumatic Stress Disorder attributable to childbirth
provides an additional impetus to increase research efforts in this area.
Alterations in maternal mood can lead to a more debilitating disease known as
Postpartum Depression (PPD). PPD is a serious health condition in child-bearing women
affecting 500,000 women and infants each year (Wisner, et al, 2002). It is the most common,
23
unrecognized complication of childbirth today. Often criticized as not an actual disease, PPD is
the primary outcome variable in many funded research studies. PPD has more recently gained
fame and notoriety in the public view via headlines and media attention (e.g., Brooke Shields,
Andrea Yates). Affective mood disorders following childbirth range in severity from early
postpartum “baby blues” to postpartum psychosis. Along this continuum lies PPD. Despite
growing knowledge, PPD screening is not common in the United States.
According to the DSM-IV, PPD is a specifier for Major Depressive Disorder (MDD) and
occurs within 4 weeks after delivery (APA, 2000). However, researchers commonly define
depression occurring within 3 months postpartum as PPD (Wisner, et al, 2002). To date, there is
not a consensus as to whether PPD represents a separate diagnosis or whether it is a variant of
MDD, Bi-Polar or minor depression. Symptoms of PPD include feelings of sadness, guilt,
inadequacy, tearfulness, hopelessness, listlessness, sleep disturbances, inability to concentrate,
and mood lability (Martinez-Schallmoser, Telleen et al. 2003).
The etiology of PPD appears to be multi-factorial and complex (Martinez-Schallmoser,
Telleen, et al, 2003). An exact cause of PPD is not known and there is little evidence to support a
biological basis (Cooper, 1998). Theories suggest that hormonal changes in pregnancy or thyroid
changes may play a role; however, there is not firm evidence to support this (Cooper, 1998).
Psychosocial risk factors and mental health history continue to consistently be linked as
the major etiological factors in PPD (Cooper, 1998). Other risk factors for PPD include a
depressive symptoms during the pregnancy, young, single, lower socio-economic status,
exposure to recent life stress, difficult infant temperament, mode of delivery, infant health at
delivery, and absence of social support (Beck, 2002). The strongest risk factors found repeatedly
in studies were depressive symptoms during pregnancy (OR 6.78) or a history of depression
24
before pregnancy (OR 3.82) (Rich-Edwards, Kleinman, et al, 2006). Socio-demographic factors
of postpartum women such as acculturation, race, ethnicity and low income have also been
studied extensively and linked to PPD (Martinez-Schallmoser, Telleen, et al, 2003). For
example, Mexican-American women with increased acculturation in the form of English spoken
as the primary language, and increased stress over extended family involvement, have been
linked to increase incidence of PPD (Martinez-Schallmoser, Telleen, et al, 2003). Similarly,
African-American and Hispanic mothers had the highest prevalence of depressive symptoms
compared to non-Hispanic whites explained primarily by low income and poor pregnancy
outcome in a study completed by Rich-Edwards ((Rich-Edwards, Kleinman, et al, 2006).
Prevalence rates of PPD vary greatly according to populations studied and instruments
used. A meta-analysis of 59 studies reported the overall prevalence of major PPD to be 13%
(Dennis, et al, 2004). Put more simply, one out of every 8 women suffers an episode of
depression after birth (Peindl, et al, 2004). Women who have suffered one episode of depression
after birth have a risk of recurrence in a subsequent birth of about 25% (Wisner, et al, 2004).
Although inception rate is greatest in the first 12 weeks, up to 50% of mothers who suffer from
PPD will remain depressed at 6 months postpartum (Dennis, et al, 2004). The most serious form
of mental illness following childbirth, postpartum psychosis, affects less than 1 % of all new
mothers (Dennis, et al, 2004). Around 60%-80% of women who give birth experience a mild
form of postpartum depression, known as the baby blues. The baby blues usually occurs within
3-10 days after birth and can last several hours or up to 10 days (Olshansky & Sereika, 2005).
Postpartum Depression negatively affects the quality of life and functional status of
mothers and infants. High levels of maternal depressive symptoms have been associated with
breastfeeding, infant interaction, infant temperament parenting skills, infant attachment, infant
25
behavioural problems and infant cognition (Beck, 2002; Chabrol, Teissedre, et al, 2002; Cooper
& Murray, 1998).
In their critical review and analysis of the literature, Grace, Evindar, and Stewart revealed
the following general conclusions:
• The strongest cognitive effects of PPD on infants were found at age 18 months;
• The strongest behavioral effects of PPD on infants were found at age 5 years as rated
by their school teachers;
• There are significant effects of PPD on latency and intensity of crying;
• Effects are enhanced with parental conflict and lower socio-economic status;
• Child gender appears to be a significant factor with boys suffering from more of the
detrimental effects (Grace, et al, 2003)
A large amount of qualitative research has also occurred in an attempt to understand the
phenomenon of PPD for the women experiencing the disorder. Mason, et al, studied the lived
experience of PPD in a psychiatric population. The subjects in this study described how their
past experience affected their thoughts and views of their labor and delivery, and their
postpartum experience (Mason, Rice, et al, 2005). Recently, PPD in women with a history of
infertility has been studied (Olshansky & Sereika, 2005). Although considered successes in the
medical community, infertile women also can suffer from PPD and often are afraid or ashamed
to admit their feelings or voice their concerns.
Psychotherapy and more recently drug therapy has been widely used as interventions in
research studies as both the treatment and as prevention of PPD. Counseling interventions are
gaining more popularity in the immediate period following delivery; however, they are not
26
proving to be very effective. Counseling interventions longer into the postpartum period have
recorded more successes (Chabrol, Teissedre, et al, 2002).
The two most widely studied groups of antidepressants used for PPD are tricyclics and
SSRIs. In a study done by Wisner, etal, in 2001, the tricyclic Nortriptyline was found to be no
more effective than the placebo in preventing a recurrence or the time to recurrence in women
with a history of PPD (Wisner, Perel, et al, 2001). Similarly, in 2004, Wisner, et al, reported that
the SSRI sertraline was more effective than the placebo in preventing recurrence and also
reported a longer time to recurrence (Wisner, Perel, et al, 2004). Most recently, Wisner, et al,
completed a study comparing Sertraline and Nortriptyline and found that psychosocial
functioning improved similarly in both groups (Wisner, et al, 2006).
Early screening and treatment of PPD has been a critical focus of the recent literature. In
fact, perinatal depression has been identified as a major public health issue by the Agency for
Heathcare Research and Quality (AHRQ). Recently, several United States senators have
introduced the MOTHERS (Mom’s Opportunity to Access Help, Education, Research, and
Support for Postpartum Depression) Act to Congress (www.ahrq.gov, 2007). The MOTHERS
Act will ensure that new mothers and their families are educated about postpartum depression,
screened for symptoms, and provided with essential services for treatment. In addition, the
MOTHERS Act will increase research into the causes, diagnoses and treatments for postpartum
depression. The MOTHERS Act was introduced in response to a recently passed New Jersey
state bill requiring physicians and nurses to educate and screen new mothers on PPD. Research
on early identification of risk factors is timely in that it will add to the literature on what is
known already, and will contribute to the body of knowledge in which health care providers
prepare their screening and education practices.
27
2.4 FUNCTIONAL STATUS
Alterations in functional status after childbirth have been linked to social support, infant care,
maternal mood and physical trauma following delivery. Functional status after childbirth is a
multidimensional construct. Theoretical basis for the concept of functional status can be partly
attributed to the role function adaptive models of Roy’s Adaptation Model (McVeigh, 2000).
Functional status is thought to be achieved when women assume their motherhood role
responsibilities and also reorganize their lives around their infant (McVeigh, 2000). Alterations
in functional status include the inability to perform tasks at the level of functioning that was
exhibited prior to childbirth. Such activities may include household chores, social and
community activities, self-care activities and occupational activities (Fawcett, 1988). Midwifery
and related nursing research have indicated that full functional status after childbirth can take up
to 6 months postpartum and typically longer then the 6 weeks allotted for recovery (McVeigh,
2000).
Many factors are thought to affect functional status and role adaptation following
childbirth including maternal age and parity. Most differences seem to disappear by 3 month
postpartum but are evident in the early postpartum period (McVeigh, 2000).
28
2.5 INFANT CARE
Infant care in the postpartum period has been studied extensively and much is known regarding
the importance of the mother-infant relationship. Infant care, although not exhibited prior to
childbirth, is a primary goal of successful functional status after childbirth. Shin, Park, and Kim
studied predictors of maternal sensitivity during the early postpartum period (Shin, et al, 2006).
Maternal sensitivity is defined as a mother’s ability to perceive and interpret her infant’s signals
and communication and then respond appropriately. It is a major factor in successful mother-
infant attachment. Through the mother-infant relationship, the mother is able to provide an
environment that is appropriate and beneficial to the infant’s growth and development. Cooper,
et al, completed a descriptive, cross-sectional study on 147 South-African women and found that
maternal sensitivity was significantly decreased in the depressed women as opposed to the non-
depressed women (Cooper, et al, 1999).
Prior studies have shown that social support influences the maternal sensitivity (Broom,
1994; Kivijarvi, et al 2004). Similarly in Shin’s, et al, study, social support has been found to be
a predictor of adequate maternal sensitivity; thus suggesting that social support is influential in
successful provision of infant care (Shin, et al, 2006).
In 1987, Zekowski, O’Hara and Willis studied the effects of maternal mood on the
mother-infant interaction (Zekowski, 1987). Their results found that the infants were sensitive to
the mother’s depressed mood and were less responsive to their mothers than were the controls.
29
Their study of 30 mother-infant dyads added to the literature regarding what is known about the
infant response to mother’s mood.
Infant care is an important outcome when considering the effects of an unplanned,
adverse labor or delivery event. It is possible that women who had negative birth experiences
may need to meet their own emotional needs before they can meet those of their infants, or may
feel obligated to meet the infant’s needs, but not their own (Kennell, et al,1979).
2.6 SIGNIFICANCE OF THE EARLY POSTPARTUM PERIOD
The postpartum period has been recognized as a time for increased vulnerability for depression
but the diagnostic criteria for PPD, particularly time of onset, have been frequently debated.
Given the timely subject of screening in the early postpartum period, the 2 week time point was
selected as the time point of interest for this study. Potential debate as to whether perinatal
depression arising in pregnancy or very early postpartum differs from late onset depression,
although there is little evidence to support this (Stowe, et al, 2004). For example, in a study
completed by Stowe, et al, they found that there were 2 distinct time points for onset of
depression, including depression occurring during pregnancy and depression beginning later than
6 weeks postpartum (Stowe, et al, 2004).
In a recent study, Heneghan, et al, suggested that pediatricians take a role in screening for
depression in the mothers of new infants (Heneghan, et al, 2000). Practical, legal and ethical
issues may come into play regarding the pediatrician’s role in assessing mothers’ mental health.
Researchers and clinicians are beginning to recommend that PPD screening occur at the neonatal
30
visit, often occurring at 2 weeks postpartum. Given that PPD is amenable to treatment,
specifically psychotherapy and pharmacotherapy, early screening and detection could result in
quicker onset of treatment and decreased morbidity for the mother and infant.
Preliminary research suggests the predictive power of maternal mood in the immediate
postpartum period (2 weeks) in the development of PPD (Dennis, et al, 2004). Significant
positive correlations have been found on measures of depressive symptomatology found at 2
weeks and time points later in the postpartum period, such as 6 weeks. Dennis, et al, used the
Edinburgh Postnatal Depression Scale to assess depressive symptomatology on a sample of 166
women at 1 week postpartum (Dennis, et al, 2004). They found that almost 30% of new mothers
in their sample exhibited depressive symptomatology at 1 week postpartum, implying that
women can experience some form of mental distress in the immediate postpartum period.
2.7 SUMMARY
There are many types of unplanned, adverse events that can occur during childbirth that can
result in a negative birth experience for the mother. The literature lends support that adverse birth
events, particularly unplanned cesarean sections or assisted vaginal births, can have detrimental
physical and mental health effects for the mother. Certain characteristics of the mother, such as
depression during pregnancy, education level or socio-economic status of the mother, maternal
age, and parity, can also affect maternal outcomes. In order to better understand the relationship
between adverse birth events and the specified outcomes, these characteristics were statistically
controlled for in this study.
31
Social support has been an extensively studied construct in many examples of life
transitions. Childbirth is such a transition. Research has shown that the amount and adequacy of
social support has positively influenced maternal outcomes following childbirth. As such, social
support was selected for use as a moderator for this study due to the assumed influential effects it
may have on the outcome variables.
Maternal mood, functional status, and infant care are all critical components to a
mother’s mental and physical health following childbirth. The literature suggests that all three of
these outcomes can be influenced by a multitude of factors. It would appear logical to assume
that an unplanned, adverse birth event could influence one or all of these outcomes. The early
postpartum period, specifically 2 weeks, has not been extensively studied with regard to the
effects of adverse birth events on maternal mood, functional status and infant care. The 2 week
time point can be viewed as a potentially critical time for postpartum screening for both maternal
and infant wellness. As focus continues to heighten on the high incidence of undiscovered
alterations in maternal mood and their potentially deadly results, early identification and its
associated predictors appeared to be an important and timely subject for investigation.
32
3.0 METHODS
3.1 DESIGN
In order to examine the relationship between adverse birth events and maternal mood, functional
status and infant care, a secondary analysis of data from Antidepressant Use During Pregnancy
(ADUP; MH R01 60335, PI, Dr. Katherine Wisner) was performed. This ongoing study enrolls
women for the investigation of anti-depressant use in pregnancy.
Secondary data analysis was an appropriate and cost effective approach for examining
these relationships. The secondary data analysis and parent study, ADUP, shared variables of
interest, specifically the identification of the actual labor and delivery events and the
identification of multiple, specific maternal outcomes. The parent study possessed recent, stable
data for addressing the secondary study’s aims.
Antidepressant Use During Pregnancy (ADUP; MH R01 60335) was developed from
recommendations by the American Psychiatric Association (APA) Committee on Research on
Psychiatric Treatments and its reports on treatment of major depressive disorder (MDD) during
pregnancy. The objective of ADUP is to study if the antidepressants to which a pregnant woman
is exposed: treat MDD, affects infant development, and serves as an additional drug exposure to
the infant.
33
More specifically, the APA work group identified areas in which data are needed:
1) whether minor physical anomalies occur with greater frequency in drug-exposed infants, 2) if
maternal weight gain and infant birth weight are compromised by drug exposure, 3) assessment
of longer-term behavioral teratogenicity in offspring, and, 4) management strategies to reduce
neonatal toxicity. The serotonin reuptake inhibitors (SSRI) are being studied because these
agents are first-line treatments for MDD. ADUP is a prospective investigation of the effects of
treatment of pregnant women with SSRIs. The mothers and their children are followed at
various checkpoints during the period of 24 months post-birth, including 2 weeks postpartum
which was this secondary analysis’s time point of interest. ADUP study participants also
included a large control group of women not being treated for depression, and not using any
antidepressant medications. Although birth data are obtained for the parent study using the
Peripartum Events Scale (PES), there was no specific plan in the parent study for analysis based
on the incidence of a negative birth event.
3.2 SAMPLE
Participants consisted of 123 women with postpartum data available for analysis from the
parent study. Inclusion criteria for ADUP are age 18 years or older, pregnant, at least 20
weeks gestation at time of entry, English-speaking, and able to provide informed consent.
Exclusion criteria are active substance abuse within the past 6 months and no obstetrical care.
Data from consented participants in the parent study were included in this secondary analysis
if the following instruments were completed after the 2 week postpartum visit: Peripartum
34
Events Scale (PES), Postpartum Support Questionnaire (PSQ), Edinburgh Postnatal
Depression Scale (EPDS), Inventory of Functional Status After Childbirth (IFSAC), and
Infant Care Survey (ICS). Participants were excluded from this study if their scheduled 2
week postpartum visit exceeded a timeframe of 4 weeks postpartum. If a participant had
more than one child entered into ADUP, the first sibling was routinely included in the study
and the second sibling was eliminated in order to maintain the independence of subjects.
Listwise deletion was selected to handle missing data after it was statistically
determined that the groups containing complete data sets and the groups with missing data
did not differ in terms of demographic characteristics and mean variable scores. Also, a
listwise deletion method still allowed sufficient power to test the study hypotheses.
Given the fixed sample size of 123, the minimum detectable effect size in terms of R2
was estimated using PASS (version 2005, NCSS, Kaysville, UT). Because prior studies did not
use the same covariates when estimating the effects of the addition of an adverse event in labor
or delivery on maternal outcomes, a range of R2 values attributed to the covariates was used for
predicting power. For the primary aim, a sample size of 123 had 80% power to detect a change in
R2 ranging from .045 to .054 attributed to the addition of the adverse events indicator variable.
This estimation included controlling for all covariates, with an estimated R2 ranging from .1 to
.25, when using an F-test from a hierarchical regression model with a significance level of .05
For the secondary aim, a sample size of 123 had sufficient power (at least 80%) to detect
a change in the R2 statistic ranging from .045 to .06 attributed to the interaction term for social
support with the adverse birth event indicator variable when using an F-test with a hierarchical
regression model and a significance level of .05. This model also assumed controlling for the
main effect terms for social support as well as all covariates.
35
3.3 VARIABLES
The variables for this study included: adverse birth event, social support, maternal age, parity,
education level, history of depression in pregnancy, antidepressant use at delivery, maternal
mood, maternal functional status, and infant care. The independent variable of interest was the
presence or absence of an adverse birth event. Social support was measured as a moderating
variable. The presence of social support may influence a woman’s response to an adverse birth
event as expressed in mood, functional status and infant care. The dependent variables in this
study were maternal mood, maternal functional status, and infant care. The covariates that were
expected to influence the outcome variables were maternal age, years of education, parity,
antidepressant use at delivery, and depression in pregnancy.
3.4 MEASURES
The measures proposed for this study were selected from those available through ADUP and
were consistent with the framework for this study. All instruments were well established with
moderate to strong psychometric properties. The covariates were collected from a series of
interviews and demographic questionnaires developed for ADUP. Interviews and questionnaires
were administered by the ADUP investigators during the participants’ initial visit upon entry into
36
the parent study. Data were retrieved, coded, and entered into SPSS by members of the ADUP
data analysis team lead by the project director, Barbara Hanusa, Ph.D.
The presence of an adverse birth event was the independent variable for this study.
Subjects having an adverse birth event were identified by positive responses and comments in
certain categories of the Peripartum Events Scale (PES). The PES is a 14-item scale developed to
quantify stressful events related to delivery (O’Hara, 1986). Data were entered onto the PES for
the parent study via a retrospective medical record review by obstetrical content experts,
specifically an OB/GYN physician and a labor and delivery nurse. For intended use, increased
scores, or responses, on the PES indicate increased stressful events related to the delivery. The
total PES score that was obtained for the parent study was not used for this study.
Specific categories on the PES that were reviewed were: precipitous labor, secondary
arrest of labor, other traumatic or life threatening events such as abruption or cord prolapse,
midforceps or vaginal breech delivery, vacuum extraction, primigravida labor longer than 20
hours, multigravida labor greater than 14 hours, primigravida second stage longer than 2 hours,
multigravida longer than 11/2 hours, abnormal FHR, abnormal contractions, fetal blood
sampling, abnormal fetal monitoring, significant lacerations, or term infant to NICU–unplanned.
In addition to a yes/no response on the PES, all categories had an additional field for comments.
For this secondary analysis, a careful review of the above listed categories and their
corresponding comments was completed on all women identified as having complete data as
defined for this investigation. A judgment was determined as to whether or not the birth had an
unplanned, adverse event. All judgments were validated by a second obstetrical content expert.
After consensus on identification of the occurrence of an adverse event between both reviewers,
participants were coded either [0] indicating no adverse event, or [1] indicating the occurrence of
37
an adverse, unplanned event(s) in labor or delivery. A participant could be identified as having
more than one adverse event as events were not mutually exclusive. After the participants were
identified as having an adverse, unplanned event(s), they were entered into new categories of
events that were created for grouping analysis in this study (Table 1).
38
Table 1. Adverse, Unplanned Events Identified from the Peripartum Events Scale
Vacuum extraction
Cesarean Section due to failure to progress, cephalo-pelvic disproportion, or failure to descend
Significant lacerations or blood loss
Term infant to NICU – Unplanned
Emergency Cesarean Section due to placenta abruption or fetal distress
Prolonged second stage of labor
Forceps delivery
Amniofusion for meconium or abnormal fetal heart rate monitoring
Precipitous delivery
Prolonged first stage of labor
Development of preeclampsia in labor
Cord prolapsed
Shoulder dystocia
Other
39
Social support, the moderating variable, was measured by the Postpartum Support
Questionnaire (PSQ). The PSQ is a self-report, 34-item likert-scale used to measure postpartum
social support in four categories (informational, material, emotional, and comparison). Each of
the 34 items has a response format with 8 options (from not important to very important and
from no support to a lot of support). In addition to a total score, categories can be summed
separately for importance and support ratings, and each category total score can range from 0 to
238 with a possible overall score of 952. Higher scores on the PSQ indicate higher levels of
postpartum social support. In five studies of postpartum women, Cronbach’s alpha scores have
ranged from .88 to .95. Test-retest reliability and concurrent validity have been established
(Logsdon, et al, 1996). All 34 items (and all categories) were included for analysis in this study.
Maternal mood, the first outcome variable, was measured by the Edinburgh Postnatal
Depression Scale (EPDS). The EPDS is a widely used, 10-item self-report likert-scale used for
the detection of postnatal depression. Each item receives a score of 0 (“No, never”) to 3 (“Yes,
most of the time”), with the higher numbers representing increased severity of a particular
symptom (Cox, et al, 1987). The total score was determined by adding all item scores, thus the
range of total score is 0 to 30. Boyd, et al, (2005) completed a brief review of all postpartum
depression screening instruments and found that across all studies, the EPDS had internal
consistency estimates ranging from .73-.87 and test-retest reliability ranging from .53-.74 (time
points ranging from 3 to 12 weeks) (Boyd, et al, 2005). A recommended cut-off point of 9/10
has been suggested to increase sensitivity for the purposes of community screening (Dennis, et
al, 2004).
Postpartum functional status of the mothers, the second outcome variable, was measured
by the Inventory of Functional Status after Childbirth (IFSAC). The IFSAC is a self-report, 36-
40
item, likert-scale designed to measure functional status after childbirth. Items are rated on 4-
point scales, arranged in 5 subscales encompassing the five dimensions of social aspects of
recovery (infant care, household activities, social and community activities, self-care activities
and occupational activities) (Fawcett, et al, 1988). A subscale mean was determined for each of
the five subscales based on the items answered within that subscale. The total IFSAC score was
determined in the same manner, using all items that are answered. The possible range can be
expanded due to presence of other children in the home, employment, or attending school (44
additional points, 16 additional points, and 20 additional points respectively). All possible
categories were used in this study. The higher the score on the IFSAC or subscales indicates
greater functional status. The possible range of total score is 0-212. Prior psychometric testing
has found that test-retest reliability was greater than 0.82 for all subscales, except the
occupational activities subscale, due to the small number of women from the study sample who
had returned to work at the time of the psychometric property testing (Fawcett, et al, 1988).
Infant care, the third outcome variable, was measured by the total score of the Infant Care
Survey (ICS). The ICS is a self-report, 52-item scale measuring parents’ beliefs about their
knowledge of and skills with infant care activities relating to health, safety, and diet on a 5-point
likert- scale ranging from 1, a great deal of confidence to 5, very little confidence (Froman, et al
1989). The total score was determined by adding all item scores. Possible range of total score is
52 to 260. Higher scores indicate less confidence regarding infant care activities. Prior studies
have found internal consistency estimates for knowledge and skills to be .95 and .96 (Froman, et
al, 1989).
41
3.5 DATA MANAGEMENT
Data were entered by the staff at Women’s Behavioral Health CARE, Pittsburgh, PA, the project
site of the parent study and reviewed by the statistician, Barbara Hanusa, PhD. The data were de-
identified by an honest broker at the parent site prior to releasing the data for analysis in this
study. Approval to conduct this study was obtained from the University of Pittsburgh IRB. The
database system at Women’s Behavioral Health CARE resides on a secure server that is behind
the University of Pittsburgh Medical Center firewall. Assessments were tracked at the parent site
to identify missing forms and generate timely reports of enrollees, where the participants were in
the study, and the anticipated time for the next interview. Data were stored by subject
identification number. Subject names and contact information were kept in separate databases
from the assessments. Paper files, such as consent forms and medical records, were kept in
locked file drawers in study coordinators offices.
3.6 ANALYSIS
Statistical guidance for the analyses in this proposal was provided by Dr. Susan Sereika,
Associate Professor at the School of Nursing and Departments of Biostatistics and Epidemiology
in the Graduate School of Public Health. Data was analyzed using SPSS (version 13, SPSS, Inc.,
42
Chicago, IL). Detailed descriptive analysis of the data was performed. Continuous variables
were analyzed for mean, range, standard deviation, skewness, kurtosis and frequency statistics.
Categorical variables were evaluated for distribution using frequency statistics and percentages.
If categories resulted in illogical or inconsistent groupings with insufficient frequencies,
variables were recoded to present more meaningful groupings. Relationships between discrete
categorical variables were examined via a two-way contingency table using the chi-square test of
independence, or by using multi-way frequency analysis among three or more discrete variables.
Case processing summaries produced in SPSS were also evaluated.
All underlying assumptions required for multiple linear regression techniques were
performed including tests for univariate and bivariate normality and sample distributions,
distribution of the residuals (estimates of the conceptual model errors), linearity, and
homogeneity of error variance. The data collected for each participant were assumed to be
independent. Statistical tests (such as Shapiro-Wilkes test for normality and Levene test for
homogeneity) as well as graphical and descriptive techniques described in the previous section
were employed to examine desired assumptions. Exploratory graphs of the continuous variables
included: histograms, normal Q-Q plots, detrended normal Q-Q plots, stem and leaf plots, and
scatter plots. Residual plots such as standardized residuals versus predicted values were also
examined. Multicollinearity among the predictor variables was also examined via estimates for
condition indices, variance decomposition proportions, tolerance, and variance inflation factors.
If the underlying assumptions were violated, data transformations on the continuous
variables (such as log or square root) were considered and assessed for use in the analysis. The
choice of transformation depended on the degree to which the sample distribution deviated from
the normal distribution. Constants were added if the original distribution contained values less
43
than one to avoid taking a log, square root, or inverse of zero. All assumptions were re-evaluated
following transformation. If transformation of a continuous variable did not appear to be helpful,
transformation of the variable into a categorical variable, using clinically meaningful cut-points,
was employed as a more suitable remedy.
This exploratory investigation also included the examination of the relationships between
outliers and dependent variables, and the independent variable. All outliers were evaluated for
validity. A sensitivity analysis excluding the outliers was performed. Outliers were identified via
statistical techniques such as frequency statistics, measures of Mahalanobis distance, and
examination of z scores, as well as graphical techniques such as bivariate plots between pairs of
variables. Once identified, steps were taken statistically to reduce their possible influence.
Patterns of missing data were evaluated. The randomness of missing data between
subjects and within a given subject was investigated using available information on subject
characteristics as well as Little’s MCAR Test to evaluate if data was missing at random. Missing
Value Analysis in SPSS was performed in order to ascertain the statistical influence, if any, the
missing data had on the variables in this design. Amount and percentage of missing data for each
variable was carefully explored. The need for potential imputation techniques was carefully
evaluated. Possible imputation techniques might have included mean imputation where the
sample mean is used to replace missing data, or regression analysis where missing data are
estimated using other variables in the sample as predictors for the missing variable.
The primary specific aim, to examine the relationship between unplanned, adverse birth
events during labor or delivery and maternal mood, maternal functional status, and infant care at
2 weeks postpartum, was investigated first. The initial review of the data was an evaluation of
the Peripartum Events Scale (PES). An Event Indicator was computed for each subject in SPSS
44
to indicate whether or not the subject had an adverse, unplanned event in labor or delivery [0=
No, 1= Yes]. Data for the outcome variables of maternal mood, maternal functional status and
infant care were used from measurements taken at the 2 week postpartum visit.
To investigate whether a relationship existed between the adverse birth events and the
outcomes of maternal mood, maternal functional status, and infant care, statistical modeling
using multiple linear regression analysis was used. The covariates: maternal age, parity,
education level, history of depression in pregnancy, and antidepressant use at delivery, was
included in the statistical model. The level of statistical significance was set at .05.
Following initial estimates of simple, unadjusted regression, adjusted estimates were
obtained by fitting the outcome-specific multiple regression models in a hierarchical manner.
Covariates were included in the first block of predictor variables, followed by the adverse event
indicator variable in the second predictor variable block. Model summary statistics to evaluate
overall fit and prediction capabilities included R2, adjusted R2, the standard error of the estimate,
and PRESS, the sum of the prediction residuals.
For the hierarchical regression, the change in the R2 statistic with the addition of adverse
event indicator variable was computed. The crude and adjusted estimated regression coefficients
and their 99% confidence intervals were obtained as effect size estimates for the adverse event
indicator variable. Residual analyses were conducted for each model estimated to check for
outliers and influential observations and to assess the distributional properties of residuals.
The secondary aim was to explore if social support was a moderator in the relationship
between unplanned birth events and maternal mood, functional status, and infant care. Social
support was measured using data received from the Postpartum Support Questionnaire (PSQ)
completed at the 2 week postpartum visit. An analytic approach similar to that described for the
45
analysis for the primary specific aim was employed. The hierarchical regression models
developed for each outcome variable in the primary aim were expanded to also include the main
effect for social support (as measured by the PSQ total scale score) in the second block of
predictor variables. The third block of predictor variables consisted of an interaction term,
computed as the product of the PSQ total scale score with the adverse events indicator variable.
To estimate moderation effects for social support on the relationship between the identified
outcome variables and adverse birth events, I examined the change in R2 statistic with the
addition of the interaction term for social support with adverse birth events to the main effects
linear regression model. The adjusted regression coefficient with confidence interval was also
obtained as an effect size estimate. Also exploratory in nature, an interaction term was created
and assessed for model influence for all of the covariates listed in the study.
46
4.0 PILOT STUDY
4.1 RELIABILITY AND VALDITY ESTIMATES FOR THE EDINBURGH
POSTNATAL DEPRESSION SCALE USED AT 2 WEEKS POSTPARTUM
4.1.1 PURPOSE
The Edinburgh Postnatal Depression Scale (EPDS), is the most widely used screening
questionnaire for PPD (Boyd, Le, & Somberg, 2005). Developed by Cox, Holden, and Sagovsky
in 1987, the EPDS was designed to assess postpartum symptoms in new mothers. The items
measure emotional and cognitive symptoms of PPD. The EPDS is a screening tool and is not
designed to diagnose MDD (Dennis, 2004). The EPDS has been used internationally and has
well-documented reliability and validity in multiple languages (Dennis, 2004). The EPDS is
typically selected for use due to its ease of use, uncomplicated interpretation, and high maternal
acceptance (Dennis, 2004). The purpose of this study was to analyze the reliability and validity
estimates of the EPDS at measures taken at 2 weeks postpartum. Although previous
psychometric studies have been conducted on this instrument, this study was specifically
examining data collected early in the postpartum period. Typically, a diagnosis of PPD occurs
later in the postpartum period; however, by hypothesizing a relationship between the symptoms
of postpartum mood and an adverse event in labor and delivery, and for the purposes of earlier
community screening, it appears to be logical and necessary to examine the psychometrics of the
47
EPDS at 2 weeks postpartum.
An exploratory aim of this pilot study was to see if the reliability estimates specific to
the data collected at the 2 week postpartum time point were consistent with reliability estimates
found on data collected during pregnancy, and on data collected during later postpartum time
points.
4.1.2 METHODS
This pilot study was a secondary data analysis for the study Antidepressant Use During
Pregnancy (ADUP; MH R01 60335) to conduct reliability and validity estimates on the EPDS at
2 weeks postpartum. A total of 229 participants were drawn from ADUP. Participants were
chosen for this pilot study if they had completed the 2 week postpartum visit. The demographic
characteristics of the sample are summarized in Table 2. The participants were primarily white,
married women with a mean age of 30.5 years (range 16-43).
48
Table 2. Participant Demographics for Pilot Study (Total n=229)
______________________________________________________________________________
Demographic Characteristics N %
______________________________________________________________________________
Race
White 181 79
African American 39 17
Other 9 4
Marital Status
Married 161 70
Unmarried 68 30
Mean SD Range
Age 30.5 5.8 16-43
Years of Education 14.8 2.1 10-17
The use of the ADUP data for a secondary data analysis was approved by the Institutional
Review Board. All data were provided by the honest broker at the parent study site and were de-
identified prior to receipt. Data were analyzed using SPSS (version 13, SPSS, Inc., Chicago, IL).
Data from consented participants in the parent study were included in this study if the following
49
instruments were completed during the defined study timeframes: 1.) EPDS, 2.) Hamilton Rating
Scale for Depression (17-item), 3.) Medical Outcomes Survey Short Form -12 (SF-12).
Measures:
Edinburgh Postpartum Depression Scale - The EPDS is a 10-item self-report likert-scale. Each
item receives a score of 0 (“No, never”) to 3 (“Yes, most of the time”), with the higher numbers
representing increased severity of a particular symptom. Responses for each of the 10 items are
scored to produce a total score for each mother. Possible total range of scores is 0-30. A cut-off
point of 12/13 at 6 weeks postpartum has a sensitivity of 68-95% when compared to a diagnosis
of major PPD established through psychiatric interview (Dennis, 2004). A recommended cut-off
point of 9/10 has been suggested to increase sensitivity for the purposes of community screening
(Dennis, 2004).
Hamilton Rating Scale for Depression (HRSD) (17-item) – The HRSD is the most frequently
used measure of outcome in antidepressant efficacy trials (Zimmerman, Posternak, &
Chelminski, 2005). The scale includes core features of depression such as low mood and loss of
interest in usual activities. Most, but not all, of the 17 items on the scale assess DSM-IV
diagnostic criteria for depression. Historically, a cut-off score of less than or equal to 7 indicates
the absence of depression and better quality of life (Zimmerman, Posternak, & Chelminski,
2005). Similar to other reliability estimates from other studies using the HRSD 17-item scale,
Zimmerman et al reported an intra-class coefficient of reliability of .97 (Zimmerman, Posternak,
& Chelminski, 2005). The HRSD -17 item consists of 17 items, nine of which are scored on a
five-category Likert-scale (from 0 to 4) and the remaining eight on a three category Likert-scale
(from 0 to 2). Total theoretical score range is from 0 to 52 (Light, et al, 2005).The HRSD 17-
50
item was selected for the validity analysis for this study because of its ease of use, its history of
strong psychometric properties, and its sensitivity for measuring similar depressive symptoms to
that of the EPDS.
Medical Outcomes Study SF-12 (Physical Component) - The MOS SF-12 is a measure of
general health status developed to compare the impact of different diseases and conditions on
health, or to monitor the health of individuals or groups over time. The MOS SF-12 contains a
subset of 12 items from the original MOS measure, the MOS SF-36, including one or two items
from each of the eight MOS SF-36 scales (Brazier & Roberts, 2004). The original MOS SF-36
has a long history of strong psychometric performance (Brazier & Roberts, 2004). The MOS SF-
12 is scored to produce two summary scores, the physical and mental health summaries (PCS
and MCS). In cross-validation with data from the Medical Outcomes Study, the PCS-36 and
PCS-12 were highly correlated 0.95 (Brazier & Roberts, 2004). Scores on the PCS range from 0
to 100, with a population mean of 50. Higher scores reflect better physical functioning. Physical
functioning is assessed on theSF-12 with items such as ‘‘Does your health limityou in climbing
several flights of stairs?’’, and ‘‘Have you accomplished less than you would like as a result of
your physical health?’’ (Katz et al, 2005). The MOS SF-12 (PCS) was selected as a measure for
discriminant validity for this study because of its ease of use, its history of strong psychometric
performance, and its contrasting assessment of physical data as opposed to the EPDS’ mental
assessment.
Two weeks postpartum was selected as the time point of interest because of the potential
effects of adverse, unplanned labor and delivery events. A negative response to such adverse
events may result in altered maternal mood early in the postpartum period. Alterations in
maternal mood can lead to PPD.
51
A detailed descriptive analysis of the data was performed. Frequencies and descriptive
statistics were analysed on the following data groups: 1.) EPDS scores at all pregnant and
postpartum time points in aggregate, 2.) EPDS scores for each of the pregnancy time points (20,
30, 36, 42 weeks gestations), 3.) EPDS scores for each of the postpartum time points (2, 12, 26
and 52 weeks postpartum), 4.) EPDS scores only at 2 weeks postpartum, 5.) HRSD-17 scores
only for 2 weeks postpartum, 6.) MOS SF-12 (PCS) scores only at 2 weeks postpartum. The
Cronbach formula for coefficient alpha was used to assess the internal consistency for the EPDS
at all time points. Spearman’s Rho correlation coefficients were used to describe the internal
item structure of the EPDS. It was expected that the reliability assessments for this study for total
item scores, pregnancy/postpartum states and the 2 weeks postpartum time point to estimate high
reliability. It was expected that the individual item correlations be significant for this sample.
Spearman’s Rho values were used to assess the relationship between the EPDS total scores at 2
weeks postpartum and the Hamilton Rating Scale for Depression 17-item total scores at 2 weeks
postpartum in order to ascertain a measure of convergent validity. It was hypothesized that the
HRSD and EPDS scores for this sample were highly correlated. Spearman’s Rho values were
also used to assess discriminant validity by measuring the relationship between the EPDS total
scores at 2 weeks postpartum and the MOS SF-12 (PCS) at 2 weeks postpartum. It was
hypothesized that the scores from each of these measure were not significantly correlated.
52
4.1.3 RESULTS
Frequency data are described in Table 3. Of the 912 available EPDS total scores from all time
points obtained from the sample, 550 (60.3%) were scores obtained during pregnancy, and 362
(39.7 %) were scores obtained during the postpartum period. Of the postpartum scores, 156 (39.7
%) were obtained at the 2 week postpartum study visit.
53
Table 3. Frequency Data for Pilot Study
Scores % Missing Mean Range
Total # EPDS Scores
Total Sample 912 100 0 5.5 0-29
Total from Pregnancy* 550 60.3 0 6 0-29
Total from Postpartum** 362 39.7 0 4.7 0-26
Total from 2 Weeks PP 156 43 0 4.7 0-20
Total # HRSD-17 Scores at 2 Weeks Postpartum
156 na 0 5.8 0-24
Total # SF-12 (PCS) at 2 Weeks Postpartum
154 na 2 44 1-64
* Time points at 20, 30, 36, 42 weeks gestation
** Time points at 2, 12, 26, 52 weeks postpartum
______________________________________________________________________________
Reliability estimates for the EPDS were strong throughout the analysis for all time points.
Estimates given were based on standardized items and covariance analysis. Reliability of the
EPDS at 2 weeks postpartum measured by Cronbach’s alpha was .875. EPDS reliability for all
time points in both pregnancy and postpartum was .908. Reliability of pregnancy and postpartum
states separately was .908 and .885 (Table 4).
54
Table 4. EPDS Reliability Data
Standardized Alpha
______________________________________________________________________________
Pregnancy .908
Postpartum .885
2 Weeks Postpartum .875
The EPDS was shown to have strong construct validity at the 2 week postpartum time
point. The EPDS was significantly correlated with the HRSD-17 with a Spearman Rho value of
.575 (p <.001). The EPDS was mildly correlated with the MOS SF-12 (PCS) at 2 weeks
postpartum with a Spearman Rho value of -.184 (ρ = .022) (Table 5).
55
Table 5. EPDS Validity Data
______________________________________________________________________________
Rho
______________________________________________________________________________
EPDS with HRSD-17 at 2 Weeks Postpartum 0.575*
EPDS with SF-12 (PCS) at 2 Weeks Postpartum -0.184**
EPDS with HRSD-17 all Postpartum 0.571*
EPDS with SF-12 (PCS) all Postpartum -0.074
* Significant at the .01 level (2-tailed)
** Significant at the .05 level (2-tailed)
4.1.4 DISCUSSION
The purpose of this pilot study was to analyze the reliability and validity estimates of the EPDS
at measures taken at 2 weeks postpartum. An exploratory aim of this study was to see if the
reliability estimates specific to the data collected at the 2 week postpartum time point were
consistent with reliability estimates found on data collected during pregnancy, and on data
collected during all postpartum time points. Many studies have used the EPDS for measurement
of maternal mood and PPD. Prior studies have reported that the EPDS is reliable, valid, brief,
56
easy to administer and use, and highly accepted across varying cultures and socio-economic
statuses (Cox, et al, 1987). Psychometric testing completed by the instrument developers
reported that a cut- off point of 12/13 at 6-weeks postpartum resulted in sensitivity estimates of
68-95% and specificity ranging from 78-96% when compared to a diagnosis of major postpartum
depression established through psychiatric interview (Cox, et al, 1987). Boyd, et al (2005)
completed a brief review of all postpartum depression screening instruments and found that
across all studies, the EPDS had internal consistency estimates ranging from .73-.87 and test-
retest reliability ranging from .53-.74 (time points ranging from 3 to 12 weeks).
As predicted, the EPDS was shown to have strong reliability across all time points in
pregnancy and postpartum. All Inter-item correlations were significant. Spearman’s Rho values
ranged from .308 - .734 and were significant at the .01 level (2-tailed) (see Table 6). Reliability
measures were consistent, and variability in scores ranged 3 % with the lowest estimate being at
the 2 week postpartum time point. These findings are similar to prior studies that have used the
EPDS (Dennis, 2004; Boyd & Somberg, 2005).
57
Table 6. Spearman’s Rho Inter-Item Correlation for All Scores
______________________________________________________________________________
Laugh Enjoy Blame Anxious Scared Stay/top Sleep Sad Cry
Laugh
Enjoy .734
Blame .427 .405
Anxious .399 .417 .621
Scared .367 .395 .548 .701
Stay/top .480 .487 .540 .522 .477
Sleep .473 .506 .516 .490 .513 .507
Sad .549 .579 .578 .572 .550 .614 .654
Cry .488 .497 .487 .463 .459 .516 .541 .706
Harm .355 .364 .336 .308 .326 .332 .382 .386 .359
* All estimates significant at the .01 level (2-tailed)
______________________________________________________________________________
The EPDS estimate of construct validity at the 2 week postpartum time point resulted in
strong validity. As predicted, the EPDS was significantly correlated with the HRSD-17 at 2
weeks postpartum; thus, proving to be a significant indicator of convergent validity. Similarly,
discriminant validity was measured by a correlation measure between the EPDS and MOS SF-12
(PCS) scores at 2 weeks postpartum. The low value of Spearman’s Rho (-.184) was indicative of
discriminant validity. The significance of the p value, despite a low correlation, was not
58
unexpected due to the large same size.
There were several limitations of this study that require discussion. First, typical with
many self-report psychosocial instruments, the data were not normally distributed due to the
multitude of zero scores by participants indicating the absence of a particular symptom. This
resulted in an abnormal distribution of data and the need to use Spearman’s Rho for correlation
calculations as opposed to Pearson Product Moment correlations.
A second limitation of this study involved the actual time of the study visit for the 2 week
postpartum score. Although it is being recorded by the parent study as the 2 week EPDS score, a
few of the time frames for visits may have varied from anywhere between 7 and 14 days. This
would affect the validity assessment reported specifically for the 2 week postpartum time point.
Later time points may have significantly affected the EPDS, HRSD-17, and MOS SF-12 (PCS)
scores and subsequently the reliability and validity estimates.
Another limitation is the lack of a heterogeneous sample. As described, the sample from
the parent study was mostly white, married, and educated. The generalizability of use with
samples of different demographic characteristics needs to be demonstrated.
The fourth limitation to be discussed concerns the lack of an estimate of test-retest
reliability. Although the data provided for this study was longitudinal in nature, the transient
state of depression, and the length of time between time points, an estimate of test-retest
reliability for the participants was not provided. An EPDS score indicating the presence or
absence of depressive symptoms at the 2 week time point could be vastly different from a score
given at 12 weeks due to the sometimes transient and fluctuating nature of the disease,
particularly if it were following an adverse, unplanned labor and delivery event.
Finally, some may argue that specifically defining the 2 week postpartum period for use
59
in estimating the validity estimate may be too narrow. However, because of a specific planned
research trajectory of studying women following an adverse, unplanned event in labor or
delivery event, this study was relevant and could be of interest to future investigators interested
in studying depressive symptoms in the early postpartum period. In response to this argument,
data correlation estimates were given for all postpartum time points using scores from the EPDS,
HRSD-17, and MOS SF-12 (PCS), and similar estimates of reliability and validity were reported.
Future research on validity estimates may include other specific postpartum time points in order
to assess the longitudinal stability of the psychometric properties of the EPDS.
Due to the results of this analysis, it is concluded that the hypotheses for this pilot
study were met and that the EPDS is a reliable and valid instrument for use in planned research
that specifically looks at data at the 2 week postpartum time point.
60
5.0 RESULTS
The study results are organized into four sections. First are the sample characteristics. Second are
the descriptive statistics for the independent, dependent and moderating variables. Third are the
results for the specific aim. Fourth and last are the results for the secondary aim.
5.1 SAMPLE CHARACTERISTICS
There were 123 women with complete data that met the criteria for this study. The sample of
women was primarily Caucasian (80%) which is consistent with the demographics of Allegheny
County, Pennsylvania. Of the 123 women, 19 were African American and 6 were within other
racial groups. Ninety-six women were married or cohabitating with a significant other. Only 27
of the women reported themselves as single. The majority of the women had more than a high
school education with 85 of them reporting college degrees or graduate college education. The
mean age of the women was 30.5 years old. The mean number of children for the women was
about 2. Twenty-six percent of the women reported being depressed at some point during the
pregnancy and 21 women were taking antidepressants at the time of delivery. Further
information on the sample characteristics are provided in Table 7.
61
Table 7. Demographic Characteristics of Sample (n=123)
N %
Race
Caucasian 98 80
African American 19 15
Other 6 5
Marital Status
Married 89 72
Single 27 22
Cohabitate 7 6
Education Level
Less Than High School 6 5
High School Graduate 9 7
Some College 23 19
College Graduate 50 41
Post Graduate 35 28
Depressed During Pregnancy 32 26
Antidepressant Use at Delivery 21 17 ______________________________________________________________________________
62
Table 7: Demographic Characteristics of Sample (n=123) Continued
______________________________________________________________________________
Mean SD Min Max
______________________________________________________________________________
Age 30.5 5.7 18 43
Parity 1.9 1.0 1 6
5.2 DESCRIPTIVE STATISTICS
5.2.1 INDEPENDENT VARIABLE: ADVERSE EVENTS
Of the 123 women in the sample, the number of women identified as having an adverse,
unplanned event in labor or delivery was 57 (46%). The adverse, unplanned events identified for
the sample from the PES are described in Table 8. The most prevalent occurrence of an adverse
event involved either a vacuum extraction or a cesarean delivery (both at 19% each), followed by
significant lacerations or blood loss at 15%. The least common adverse events included
prolonged first stage of labor, development of preeclampsia in labor, cord prolapse, and shoulder
dystocia- all at 2% each. The occurrence of multiple events was not quantified for the purposes
63
of this study; however, it is important to note that the events listed in table are not mutually
exclusive.
Table 8. Adverse, Unplanned Events Identified in the Sample
_____________________________________________________________________________
N %
______________________________________________________________________________
Vacuum Extraction 10 19
Cesarean Section due to FTP, CPD, FTD 10 19
Significant lacerations or blood loss 8 15
Term Infant to NICU-Unplanned 6 11
Emergency C-Section due to abruption or fetal distress 6 11
Prolonged second stage of labor 5 9
Forceps delivery 3 6
Amniofusion for meconium or abnormal HR monitoring 3 6
Precipitous delivery 2 4
Prolonged first stage of labor 1 2
Development of preeclampsia in labor 1 2
Cord Prolapse 1 2
Shoulder dystocia 1 2
Total N 57 46
64
5.2.2 DEPENDENT AND MODERATING VARIABLES: MATERNAL MOOD,
FUNCTIONAL STATUS, INFANT CARE, SOCIAL SUPPORT
The mean, standard deviation, range of scores, minimum and maximum scores for the 3
dependent variables are described in Table 9. Average maternal mood score was 4.69. Over 26
(21%) of the postpartum mothers scored above or equal to the cut-point score of 9. Mean
functional status after childbirth was 69.5. The mean infant care score was 104.8. Average total
social support was 185.3. The mean social support score was 185.3.
Table 9. Measures of Central Tendency for the Independent Variables
Mean + Range Min Max SD
Maternal Mood 4.69 + 4.8 20 0 20
Functional Status 69.5 + 15.8 55.4 7 99.2
Infant Care 104.8 + 33.8 33.8 61 214
Social Support 185.3 + 86 385 3 388
65
5.3 PRIMARY AIM
For the primary aim, it was hypothesized that unplanned, adverse birth events were associated
with altered maternal mood, maternal functional status and infant care at 2 weeks postpartum. In
order to examine the relationship between adverse birth events and maternal mood, functional
status and infant care, a secondary analysis of data from the study on Antidepressant Use During
Pregnancy (ADUP; NIH R01 MH60335) was completed.
5.3.1 MATERNAL MOOD
Because the data collected on the EPDS were not normally distributed, EPDS scores were
categorized into 2 groups for use with binary logistic regression analysis. As per the
recommended guideline given for the purpose of depression screening (Dennis, 2004), total
scores ranging from 0-8 were given a value of [=0] and total scores equal to or greater than 9
were given a value of [=1]. To investigate whether a relationship existed between the adverse
birth events and maternal mood, logistic regression was used to compare each group of EPDS
scores by the adverse event incidence (no event occurrence [=0] versus any event occurrence
[=1]).
66
Upon examination of the logistic regression results, except for the covariate of
depression in pregnancy (odds ratio=6.94, 95% confidence interval: 2.11, 21.844, p=.001), an
adverse event in labor or delivery did not predict maternal mood when using the 2 categories of
scores as provided by the EPDS data (odds ratio=1.38, 95% confidence interval: 474, 3.79,
p=.536). None of the other covariates were significant in the multiple regression analysis. Further
detail regarding the logistic regression results are described in Table 10.
Table 10. Logistic Regression Results for Maternal Mood
95% C.I. for Odds Ratio
Odds Ratio Lower Upper
Depression in Pregnancy 6.79** 2.11 21.844
Adverse Event 1.34 .474 3.79
** Significant at .05
67
5.3.2 FUNCTIONAL STATUS
To investigate whether a relationship exists between the adverse birth events and the outcomes of
maternal functional status, statistical modelling using multiple linear regression analysis was
used to compare this continuous type maternal outcome by adverse event incidence (no event
occurrence [=0] versus any event occurrence [=1]). The analytic approach for both the primary
and secondary aims allowed for the inclusion and examination of potential confounding
variables, maternal age, parity, education level, depression in pregnancy, and antidepressant use
at delivery, while modelling the primary relationships of interest. The level of statistical
significance was set at .05. Residual analyses were conducted on all models to assess model fit.
Although several model misspecifications and influential observations seemed apparent, based
on sensitivity analyses, the regression results for each model did not change significantly with
the removal of any influential observations.
Based on the change in R2 using multiple linear regression, less than 1 % of additional
variance was explained in functional status (R2=.696, p=.66) following an adverse event in labor
or delivery. The aggregate effect of the covariates entered into block 1 of the hierarchical model
was significant for functional status (R2=.695, p<.001). Further detail regarding the multiple
linear regression results for functional status are described in Table 11.
68
Table 11. Multiple Linear Regression Coefficients for Functional Status
______________________________________________________________________________
R2 SE p R2 Change
Covariates* .695 9.1 <.001 na
Adverse Event Indicator .696 9.14 .66 .001
*Maternal age, parity, education level, antidepressant use at delivery, depression during
pregnancy
______________________________________________________________________________
5.3.3 INFANT CARE
To investigate whether a relationship exists between the adverse birth events and the outcomes of
infant care, statistical modeling using multiple linear regression analysis was used to compare
this continuous type maternal outcome by adverse event incidence (no event occurrence [=0]
versus any event occurrence [=1]).
Based on the change in R2 using multiple linear regression, less than 1% of additional
variance was explained in infant care (R2=.321, p=.41). The aggregate effect of the covariates
entered into block 1 of the hierarchical model was significant for functional status (R2=.317,
p<.001). Further detail regarding the multiple linear regression results for infant care are
described in Table 12.
69
Table 12. Multiple Linear Regression Coefficients for Infant Care
_____________________________________________________________________________
R2 SE p R2
Change
Covariates* .317 .116 <.001 na
Adverse Event .321 .116 .418 .004
*Maternal age, parity, education level, antidepressant use at delivery, depression during
pregnancy
5.4 SECONDARY AIM
For the secondary aim, it was hypothesized that social support directly influences the maternal
response at 2 weeks postpartum to negative birth events and may moderate the effect of
unplanned adverse events during labor/delivery. To explore if social support was a moderator in
the relationship between unplanned birth events and maternal mood, functional status, and infant
care, an analytic approach similar to that described for the analysis of data for the primary
specific aim was employed. The hierarchical regression models developed for each outcome
variable in the primary aim was expanded to also include the main effect for social support and
70
the interaction effect of the adverse event indicator with social support after controlling for the
covariates of maternal age, parity, education level, antidepressant use during delivery, and
depression during pregnancy. Residual analyses were conducted on this model as well. All
model misspecifications or influential observations were removed and the analyses were
repeated to determine if statistical results would change. In all cases, statistical results did not
change with the removal of any influential observations.
5.4.1 MAIN EFFECT OF SOCIAL SUPPORT
Social support was a weak predictor of maternal mood with a likelihood ratio of 92.14 when
added to the initial hierarchical model for logistic regression (odds ratio: 1.008, 95% confidence
interval: 1.001, 1.016, p=.02). Adding social support to the hierarchical model for multiple linear
regression was the strongest explanatory variable with an additional 1.6% of explained variance
in functional status (R2=.712, p=.014), and 9% additional explained variance in infant care
(R2=.415, p<.001).
5.4.2 MODERATING EFFECT OF SOCIAL SUPPORT
The interaction of social support and the adverse event indicator when added to the logistic
regression model to predict mood was borderline significant (odds ratio=1.015, 95% confidence
interval: .999, 1.028, p=.045). The interaction of social support and the adverse event indicator
when added to the multiple linear regression hierarchical model explained less than 1% of
71
additional variance in functional status (R2=.721, p=.056), and no additional variance in the
infant care (R2=.415, p=.92). The effect of social support was linear and linear in the logit when
added to the multiple linear regression models and logistic regression model, respectively. This
was evaluated by adding a transformed variable of social support (square root) to the hierarchical
regression models already defined. Detailed results for social support when added to the logistic
regression and multiple linear regression models are provided in Tables 13, 14 and 15.
Table 13. Logistic Regression Results for Maternal Mood and Social Support
____________________________________________________________________________
95% C.I for Odds Ratio
Odds Ratio Lower Upper
______________________________________________________________________________
Depression in Pregnancy 6.79** 2.11 21.844
Adverse Event 1.34 .474 3.79
Social Support 1.008** 1.001 1.016
Social Support*Adverse Event 1.01 .999 1.028
**Significant at .05 level
______________________________________________________________________________
72
Table 14. Multiple Linear Regression Coefficients for Functional Status and Social Support
R2 SE p R2 Change
Covariates* .695** 9.1 <.001 na
Adverse Event .696 9.14 .66 .001
Social Support .712** 8.93 .014 .016
Social Support*Adverse Event .721 8.82 .056 .009
* Maternal age, parity, education level, antidepressant use at delivery, depression during
pregnancy
**Significant at the .05 level
______________________________________________________________________________
73
Table 15. Multiple Linear Regression Coefficients for Infant Care and Social Support
_____________________________________________________________________________
R2 SE p R2 Change
______________________________________________________________________________
Covariates* .317* .116 <.001 na
Adverse Event .321 .116 .418 .004
Social Support .415* .109 <.001 .093
Social Support*Adverse Event .415 .109 .92 0
* Maternal age, parity, education level, antidepressant use at delivery, depression during
pregnancy
**Significant at the .05 level
_____________________________________________________________________________
74
6.0 DISCUSSION
The discussion section is organized into six sections. First are the statements addressing the
characteristics of the sample. Second is a discussion of the results including the following
variables: adverse birth event, functional status, maternal mood, and social support. The third
section addresses the clinical implication of my statistically significant finding. Fourth are
limitations of the study which include: outcomes, study design, and sample. Fifth are
suggestions for future research which include: timing, instruments, definition of adverse events,
and concept measurement. Sixth and last is the concluding statement.
6.1 CHARACTERISTICS OF THE SAMPLE
The women in this sample were primarily white, married and educated. After reviewing the
literature on racial, ethnic, and socio-demographic risk factors in PPD, it is evident that race and
ethnicity do not qualify as primary, consistent risk factors for PPD. Rather, it clear that the
aggregate of many psycho-social, demographic, cultural and maternal factors influence PPD.
Therefore, although the sample characteristics, which are consistent with demographic data for
Allegheny County, PA, are not diverse, the characteristics are consistent with the past literature
in this area of study.
75
6.2 DISCUSSION OF RESULTS
The purpose of this secondary analysis was to examine the relationship between unplanned,
adverse birth events during labor or delivery and maternal mood, maternal functional status, and
infant care at 2 weeks postpartum. The findings in this study did not indicate that women with
adverse, unplanned events in labor or delivery had varying outcomes related to mood, functional
status, or infant care at 2 weeks postpartum
A secondary aim was to explore if social support was a moderator for the relationship
between unplanned, adverse birth events and maternal mood, functional status, and infant care at
2 weeks postpartum. The findings in this study did not indicate that social support was a
moderator for the relationship between unplanned, adverse birth events and maternal mood,
functional status, and infant care at 2 weeks postpartum.
Although perceptions of negative birth experiences, physical health following delivery,
cesarean and assisted vaginal delivery in relation to a later diagnosis of PTSD and postpartum
depression, and the main effects of social support have all been studied, there exists a gap in the
literature specifically looking at a variety of unplanned, adverse events in labor or delivery, such
as the events listed in Table 1, and maternal outcomes in the early postpartum period (2 weeks
following delivery). When research studies have considered the question of negative birth
experiences, the general paradigm is to compare emotional reactions to cesarean sections versus
vaginal births; however, the types and circumstances surrounding a negative birth experience can
encompass many different events as described in this study.
76
6.2.1 ADVERSE BIRTH EVENTS
In a prospective, longitudinal study done by Creedy, et al, examining women (n=499) in the third
trimester and following them until 4 to 6 weeks postpartum, it was discovered that one in three
women (33%) identified a negative birth event and reported the presence of at least three trauma
symptoms following the birth of their infant (Creedy, et al, 2000). In this study, as many as 46%
of the women experienced an adverse event as identified by a retrospective medical review.
Without specific interviews with the women in this study, it is unknown whether or not they
perceived their deliveries as having an unplanned, adverse event. The acknowledgement of an
adverse event by the women, such as in Creedy’s study, could alter their self-report scores on the
scales used in this study and thus the findings could have differed significantly.
Murphy, Pope, Frost, and Liebling completed a qualitative study (n=27 women) and
determined that operative delivery had a noticeable impact on women's views about future
pregnancy and delivery (Murphy, et al, 2003). Three important conclusions were reached by the
investigators: (1) antenatal variables did not contribute to the development of acute or chronic
trauma symptoms, (2) women who experienced both a high level of obstetric intervention and
dissatisfaction with their intrapartum care were more likely to develop trauma symptoms than
women who received a high level of obstetric intervention or women who perceived their care to
be inadequate, and (3) PTSD is evident in women after a childbirth experience that include
intrusive obstetric intervention during labor and delivery, and the care provided to women in
labor. Their study also examined women during later time points in the postpartum time period.
Being a qualitative study, the investigators were able to collect richer data as it relates to the
77
women’s perceptions; therefore possibly explaining their rich qualitative findings, as opposed to
the insignificant quantitative findings in this study. In contrast, this study examined variables
early in the postpartum period and did not have the benefit of maternal qualitative interviews.
The inability to measure women’s expectations of their labors, and their subsequent perceptions
of the event following their labors, greatly impacted our ability to accurately measure the adverse
effects of certain events. Similarly, cultural and experiential differences were unknown in this
study and may have significantly impacted the womens’ personal modeling of crisis situations
and their subsequent coping strategies.
6.2.2 FUNCTIONAL STATUS
Similar to the measurement of functional status after childbirth, Lydon-Rochelle,
Holt and Martin assessed the association between method of delivery and the general health
status, sexual, bowel and urinary functioning of primiparous women as measured at 7 weeks
postpartum (Lydon-Rochelle, et al, 2001). At 7 weeks postpartum, women who had cesarean or
assisted vaginal deliveries reported significantly lower postpartum general health status scores
than women with unassisted vaginal delivery (e.g., did not use forceps or vacuum extraction).
Additionally, women with assisted vaginal deliveries reported significantly worse sexual, bowel
and urinary functioning. It is possible that a 2 week postpartum time point may not accurately
measure physical functioning, although it would seem logical that physical ailments as a result of
childbirth would be more prevalent closer to the event causing the injury. This study did not find
significant results for functional status at the 2 week time point as Lydon-Rochelle did at the 7
week time point. This study included more types of adverse events than just assisted vaginal
78
deliveries. It is possible that by only examining women from our sample with assisted vaginal
deliveries, we could have found significant results similar to Lydon-Rochelle, et al.
6.2.3 MATERNAL MOOD
Similar to Creedy, Shochet, and Horsfall’s study, which examined psychological health
following an adverse birth event, much of the focus in prior research related to maternal
outcomes following an adverse birth event is in the area of Posttraumatic Stress Disorder (PTSD)
(Creedy, et al, 2000). For instance, Creedy, et al, found that twenty-eight women in their sample
(5.6%) met DSM-IV criteria for acute PTSD (Creedy, et al, 2000). In their study, the level of
obstetric intervention experienced during childbirth (beta = 0.351, p < 0.0001) and the perception
of inadequate intrapartum care (beta = 0.319, p < 0.0001) during labor were consistently
associated with the development of acute trauma symptoms. Typically, studies investigating
PTSD focus on later periods during the postpartum recovery where symptoms of PTSD become
more evident and more debilitating and not early in the postpartum time period. In contrast, this
study did not examine specific symptomotology of PTSD, but rather more general alterations in
a larger category of maternal mood in the early postpartum period. The instrument used in this
study to measure alterations in mood was not designed, nor nearly specific enough, to measure
the mental and emotional impact of PTSD.
79
6.2.4 SOCIAL SUPPORT
Similar to prior studies examining postpartum social support, this study found that the main
effect of postpartum social support was significant when predicting mood, functional status and
infant care. It was weakly significant for predicting mood when combined with an adverse event;
however, due to the borderline significance of the statistic, and the influential main effect of
social support by itself, it is possible that this result does not bear true statistical significance. It
would logical to assume that social support following an adverse labor event would result in
substantial clinical significance. In other words, giving social support to a mother following a
traumatic labor or delivery event would most likely yield positive results.
6.3 CLINICAL SIGNIFICANCE
Due to the significant relationship between social support and maternal outcomes, careful
prenatal screening regarding postpartum social support should be conducted. By incorporating
social support screening during the labor and delivery admission, nurses can serve to improve
outcomes by providing resources for support in the event that social support is minimally
available. Although by 4 to 6 weeks postpartum, alterations in mood, functional status or infant
care may have already developed, or even subsequently resolved, postpartum visits should allow
for appropriate screening of not only outcomes development, specifically mood alterations, but
also the presence or need for additional social support. Nurses can play a key role in not only
80
assessing a mother’s well-being during or following the birth of a child, but also can contribute
to a patient’s pool of resources that provides social support.
6.4 LIMITATIONS
6.4.1 CHOSEN OUTCOMES
This study did not look at all possible maternal outcomes. For instance, maternal role
gratification and general health were not examined in this study, but are areas of potential focus
for future studies. This study also did not look at all potential moderators that could influence
maternal outcomes. Specifically, the study did not examine the woman’s understanding of what
actually happened during the course of her labor and delivery, nor were individual coping
strategies measured.
6.4.2 STUDY DESIGN
General limitations of completing a secondary data analysis must also be addressed as the
original hypothesis for the data being collected varies from the hypotheses proposed here, and
the data collection methods were defined for the original ADUP study, not for the study being
currently addressed.
81
6.4.3 SAMPLE
Another possible limitation of this study was the sample size. A larger sample could have
resulted in more statistically significant findings when considering the presence of an adverse
birth event with alterations in mood, functional status and infant care, and in the use of more
complex models when considering moderation effects.
6.5 FUTURE STUDIES
6.5.1 TIMING
Similar to prior studies, future studies may include a longitudinal study examining the effects of
adverse birth events over time. Research has shown that a diagnosis of PPD can surface between
1 and 6 months, with 3 months being the most common time point (Gaynes, et al, 2005). Most of
the women in this sample were free of depression symptoms. In fact, only 21% scored above the
cut-off point used for community screening. This study did not examine whether or not
depression symptoms developed later in the postpartum period. A possible time point of
comparison to re-analyze this data collected at 2 weeks could be 3 months postpartum.
In contrast to examining the potential effects during later time points, we could remain with our
present estimation that the greatest impact from adverse birth events is in the early postpartum
82
period. By measuring maternal mood prior to 2 weeks postpartum, or at birth, we potentially
could discover a significant difference in maternal mood between those women who did and did
not have unplanned, adverse events during childbirth.
6.5.2 INSTRUMENTS
Another future investigation could involve the instruments used for the study. Although the
findings of the pilot study reported that the EPDS is a valid and reliable instrument for use at 2
weeks postpartum, not all of the instruments selected for this study have as strong of a
psychometric history or a total score that is as simply obtained and as indicative of concept
measurement as the EPDS. Rather than using total scores for some of the other instruments
chosen for this study, we could focus on certain subscales of the outcomes measures used in this
study. For instance, rather than look at all aspects of the Infant Care Survey, an analysis could be
conducted specifically looking at the infant care activities involved with infant health. Similarly
with functional status, the data could be analyzed specifically looking at functional status as
defined for care of the infant. This approach would allow the investigator to focus on specific
activities rather than broad-based total scores of a particular construct.
In addition to looking only at certain subscales of the measures already defined for this
study, we could use entirely different scales with potentially stronger and more sustained
histories of psychometric success, and not specifically developed for postpartum women. For
instance, by using the Hamilton Rating Scale for Depression-29 item and the Medical Outcomes
Survey SF-12 Physical Component Score, both at 2 weeks postpartum, we could potentially
83
elicit a statistically significant finding when examining maternal mood and functional status
following an unplanned, adverse labor or delivery event.
6.5.3 DEFINITION OF ADVERSE EVENTS
As previously mentioned, future analyses could include only specific types of adverse events for
examination in this hierarchical model, such as forceps or vacuum extraction deliveries. Another
approach could be to quantify or rank the adverse events in order of significance. One could
easily argue that a placental abruption which resulted in an emergency cesarean section and a
term newborn to the neonatal intensive care unit would most likely result in significant
alterations in maternal mood as compared to someone who might have experienced significant
vaginal and perineal lacerations. By differentiating between not only the severity of the event,
but also by the number of events experienced, we could perhaps be able to better quantify the
detrimental effects for the mother.
6.5.4 CONCEPT MEASUREMENT
Within the healthcare community, there seems to be expert agreement that adverse birth events
exist, and can potentially have a debilitating effect on the mother. The challenge appears to be
the ability to accurately and systematically quantify the additional stress given that childbirth is
already a life-changing, stressful event by itself. The ability to measure the additive effect of
stress continues to be an opportunity for future research and instrument development. Since we
84
already know that qualitative studies have found clinical significance in this area, one other
potential avenue would be to explore a physiologic differentiator, such as cortisol levels at birth
obtained from maternal serum or infant cord blood. With the absence of an instrument with a
proven history of success for measuring the added stress of obstetrical events, a physiologic
measure might be able to bridge the gap for researchers determined to quantify the stress level of
the birth event.
6.6 CONCLUSION
Although the data used for this study met the requirements for the stated hypotheses, and the
community-dwelling sample represented a diverse population in terms of health, and labor and
delivery management, the occurrence of an adverse, unplanned event during labor or delivery did
not appear to significantly impact several maternal outcomes at 2 weeks postpartum. This study
adds to the literature by reinforcing the need for frequent, repetitive screenings of alterations or
problems with mood, functional status, or infant care. Although it may appear logical that
women with adverse events in delivery would exhibit signs of alterations early in the postpartum
period, there is little reason to believe that women with uneventful, planned deliveries may not
experience the same types of alterations; therefore, necessitating the need for consistent nursing
quality, teaching, screening, and support following the birth of an infant.
85
APPENDIX
IRB APPROVAL
University of Pittsburgh Institutional Review Board Exempt and Expedited Reviews University of Pittsburgh FWA: 00006790 University of Pittsburgh Medical Center: FWA 00006735 Children’s Hospital of Pittsburgh: FWA 00000600 3500 Fifth Avenue Suite 100 Pittsburgh, PA 15213 Phone: 412.383.1480 Fax: 412.383.1508 TO: Ms. Diane Hunker FROM: Christopher M. Ryan, PhD, Vice Chair DATE: November 7, 2006 PROTOCOL: Effects of Birth Events on Postpartum Mood and Functional Status IRB Number: 0610027 The above-referenced protocol has been reviewed by the University of Pittsburgh Institutional Review Board. Based on the information provided in the IRB protocol, this project meets all the necessary criteria for an exemption, and is hereby designated as “exempt” under section 45 CFR 46.101(b)(4). • If any modifications are made to this project, please submit an ‘exempt
86
modification’ form to the IRB. • Please advise the IRB when your project has been completed so that it may be officially terminated in the IRB database. • This research study may be audited by the University of Pittsburgh Research Conduct and Compliance Office. Approval Date: November 7, 2006
87
BIBLIOGRAPHY
Adewuya, A.O. (2006). Early postpartum mood as a risk factor for postnatal depression in Nigerian women. American Journal of Psychiatry, 163, 1435-1437. Agency for Heathcare Research and Quality (AHRQ). (2007). Retrieved May 22, 2007 from http//: www.ahrq.gov/clinic/epcsums/peridepsum.html. Algovik, M., et al. (2004). Genetic influence on dystocia. Acta Obstet Gynecol Scand, 83(9), 832-7. Allen, V.M., O'Connell, C., Baskett, T.F. (2005). Maternal and perinatal morbidity of caeserean delivery at full cervical dilatation compared with caeserean delivery in the first stage of labor. BJOG: an International Journal of Obstetrics and Gynecology, 112, 986-990. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th edition). Washington DC: Author. Ayers, S. (2004). Delivery as a traumatic event: prevalence, risk factors, and treatment for postnatal posttraumatic stress disorder. Clinical Obstetrics & Gynecology, 47(3), 552-67. Ayers,S., Pickering,A.D. (2005). Women's expectations and experience of birth. (2005). Psychology and Health, 20(1), 79-92. Bailham, D. & Joseph, S. (2003). Post-traumatic stress following childbirth: a review of the emerging literature and directions for research and practice. Psychology, Health, and Medicine, 8(2), 160-168. Beck, C. T. (2002). Theoretical perspectives of postpartum depression and their treatment implications. MCN American Journal of Maternal Child Nursing, 27(5), 282-7. Beck, C.T. (2004). Birth trauma: in the eye of the beholder. Nursing Research, 53(4), 28-35.
88
Beck, C.T. (2004). Post-traumatic stress disorder due to childbirth: the aftermath. Nursing Research, 53(4), 216-24. Beck, C. T., Froman, R.T., et al. (2005). Acculturation level and postpartum depression in Hispanic mothers. MCN American Journal of Maternal Child Nursing, 30(5), 299-304. Boyd, R. C., Le, H. N., & Somberg, R. (2005). Review of screening instruments for postpartum depression. Archives of Womens Mental Health, 8(3), 141-153. Brazier, J. E. & Roberts, J. (2004). The estimation of a preference-based measure of health from the SF-12. Medical Care, 42(9), 851-859.
Chabrol, H., Teissedre, F., Saint-Jean,M., Teisseyre,N., Roge,B., Mullet,E. (2002). Prevention and treatment of postpartum depression: a controlled randomized study on women at risk. Psychological Medicine, 32(6), 1039-1047. Chaudron,L.H., Szilagyi, P.G., Campbell, A.T., Mounts, K.O., McInerny,T.K. (2007). Legal and ethical considerations: risks and benefits of postpartum depression screening at well-child visits. Pediatrics, 119(1), 123-128. Coffman, S., Levitt, M.J., Deets, C., Quigley, K.L. (1991). Close relationships in mothers of distressed and normal newborns: support, expectancy confirmation, and maternal well-being. Journal of Family Psychology, 5 (1), 93-107. Cooper, P. J. & Murray, L. (1998). Postnatal depression. Bmj, 316(7148), 1884-6. Cooper, P.J., Tomlinson, M., Swartz, L., Woolgar, M., Murray, L., Molteno, C. (1999). Post-partum depression and the mother-infant relationship in a South African peri-urban settlement. The British Journal of Psychiatry, 175, 554-558. Cox, J.L., Holden, J.M., Sagovsky, R. (1987). Detection of postnatal depression: development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry, 150, 782-786. Creedy, D.K., Shochet, I.M., & J. Horsfall, (2000). Childbirth and the development of acute trauma symptoms: incidence and contributing factors. Birth, 27(2), 104-11. Cunningham, F. G., MacDonald, P.C., Leveno, K.J., Gant, N.F., Gilstrap, L.C. (1993). Williams Obstetrics (19th edition). Connecticut: Appleton and Lang. Dennis, C. L., Janssen, P.A., et al. (2004). Identifying women at-risk for postpartum depression in the immediate postpartum period. Acta Psychiatr Scand , 110(5), 338-46.
89
Dennis, C. L. (2004). Can we identify mothers at risk for postpartum depression in the immediate postpartum period using the Edinburgh Postnatal Depression Scale? Journal of Affective Disorders, 78(2), 163-169. Dennis, C-L. (2005). Psychosocial and psychological interventions for prevention of postnatal depression: systematic review. BMJ, http:bmj.bmjjournals.com/cgi/content/full/331/7507/15. Retrieved on March 17, 2006. Fawcett, J., Tulman, L. & Myers, S.T. (1988). Development of the inventory of functional status after childbirth. Journal of Nurse Midwifery,33(6), 252-60. Fawcett, J., et al. (2005). Adaptation to cesarean birth: implementation of an international multisite study. Nursing Science Quarterly,18(3), 204-10. Froman, R.D. & Owen, S.V. (1989). Infant care self-efficacy. Sch Inq Nurs Pract, 3(3), 199-211; discussion 213-5. Gardella, C., Taylor, M., Benedetti, T., Hitti, J., Critchlow, C. (2001). Presented at the 21st annual Meeting of the Society for Maternal-Fetal Medicine, Reno, Nevada, February 5-10,2001. Gaynes, B.N., et al. (2005). Perinatal depression: prevalence, screening accuracy, and screening outcomes. Evid Rep Technol Assess (Summ), (119), 1-8. Grace, S.L., Evindar, A., Stewart, D.E. (2003). The effect of postpartum depression on child cognitive development and behavior: a review and critical analysis of the literature. Archives of Womens Mental Health, 6, 263-274. Hamilton, M. (1980). Rating depressive patients. Journal of Clinical Psychiatry, 41(12 Pt 2), 21-4. Heneghan, A.M., Mercer, M., Deleone, N.L. (2004). Will mothers discuss parenting stress and depressive symptoms with their child's pediatrician? Pediatrics, 113, 460-467. Horowitz, J.A., Goodman, J.H. (2004). Identifying and treating postpartum depression. JOGNN, 34(2), 264-273. Horowitz, J.A., Cousins, A. (2006). Postpartum depression treatment rates for at-risk women. Nursing Research, 55(2), 23-27. Hudelist, G., et al. (2005). Factors predicting severe perineal trauma during childbirth: role of forceps delivery routinely combined with mediolateral episiotomy. American Journal of Obstetrics Gynecology, 192(3), 875-81. Hupcey, J.E. (1998). Clarifying the social support theory-research linkage. Journal of Advanced Nursing, 27(6), 1231-41.
90
Kennell, J.H. & Klaus, M.H. (1979). Early mother-infant contact: effects on the mother and the infant. Bull Menninger Clin, 43(1), 69-78. Klaus, M. (1998). Mother and Infant: Early Emotional Ties. Pediatrics, 102(5), 1244-1246. Koo, V., Lynch, J., & Cooper, S. (2003). Risk of postnatal depression after emergency delivery. J. Obstet. Gynaecol. 29(4), 246-250. Logsdon, M.C., McBride, A.B., & Birkimer, J.C. (1994). Social support and postpartum depression. Res Nurs Health, 17(6), 449-57.
Logsdon, M.C., et al. (1996). The Postpartum Support Questionnaire: reliability and validity. Journal of Nursing Measures, 4(2), 129-42.
Logsdon, M.C. & Davis, D.W. (2003) Social and Professional support for pregnant and parenting women. MCN The American Journal of Maternal/Child Nursing, 28(6), 371-376. Lydon-Rochelle, M.T., Holt, V.L. & Martin, D.P. (2001). Delivery method and self-reported postpartum general health status among primiparous women. Paediatr Perinat Epidemiol, 15(3), 232-40. Lydon-Rochelle, M.T., Holt, V.L., Martin, D.P., Easterling, T.R. (2000). JAMA, 283(18), 2411-2416. Martinez-Schallmoser, L., Telleen, S., et al. (2003). The effect of social support and acculturation on postpartum depression in Mexican American women. Journal of Transcultural Nursing, 14(4), 329-38. Mason, W.A., Rice, M.J., Records, K. (2005). The lived experience of postpartum depression in a psychiatric population. Perspectives in Psychiatric Care, 41(2)52-61. Matza, L.S., et al. (2003). Depression with atypical features in the National Comorbidity Survey: classification, description, and consequences. Archives of General Psychiatry, 60(8), 817-26. McVeigh, C. (2000). Investigating the relationship between satisfaction with social support and functional status after childbirth. MCN The American Journal of Maternal/Child Nursing, 25(1), 25-30. Morley, W.E., Messick,J.M., & Aguilera,D.C. (1967). Crisis: paradigms of intervention. Journal of Psychiatric Nurses in Mental Health Service, 5(6), 531-44.
91
Murphy, D.J., et al. (2003). Women's views on the impact of operative delivery in the second stage of labour: qualitative interview study. Bmj, 327(7424), 1132. National Vital Statistics Reports, (2006). 55(1), 65. O'Hara, M.W. (1986). Social support, life events, and depression during pregnancy and the puerperium. Archives in General Psychiatry, 43(6), 569-73. O'Hara, M.W., Varner, M.W., Johnson, S.R. (1986). Peripartum Events Scale. Journal of Reproductive and Infant Psychology, 4, 85-98. Olshansky, E. & Sereika,S. (2005). The transition from pregnancy to postpartum in previously infertile women: a focus on depression. Archives in Psychiatric Nursing 19(6), 273-80. Patel, R. R., Murphy, D. J., & Peters, T. J. (2005). Operative delivery and postnatal depression: a cohort study. BMJ, http//:doi:10.1136/bmj.38376.603426.d3 (published 25 February 2005). Retrieved November 28, 2006. Peindl, K. S., Wisner, K. L.,et al. (2004). Identifying depression in the first postpartum year: guidelines for office-based screening and referral. Journal of Affective Disorders, 80(1), 37-44. Rich-Edwards, J. W., Kleinman, K. et al. (2006). Sociodemographic predictors of antenatal and postpartum depressive symptoms among women in a medical group practice. Journal of Epidemiology & Community Health, 60(3), 221-7. Shin, H., Park, Y.J., Kim, M.J. (2006). Predictors of maternal sensitivity during the early postpartum period. Journal of Advanced Nursing, 55(4), 425-434. Simkin, P. (1992). Just another day in a woman's life? Part II: Nature and consistency of women's long-term memories of their first birth experiences. Birth, 19(2), 64-81. Soet, J.E., Brack, G.A., & DiIorio, C. (2003). Prevalence and predictors of women's experience of psychological trauma during childbirth. Birth, 30(1), 36-46. Stowe, Z.N., Hostetter, A.L., Newport, D.J. (2004). The onset of postpartum depression: implications for clinical screening in obstetrical and primary care. http://www.mdconsult.com/das/article/body/69708238-4/jorg. Retrieved on April 30, 2007. Thoppil, J. , Riutcel, T.L., & Nalesnik, S.W. (2005). Early intervention for perinatal depression. American Journal of Obstetrics & Gynecology, 192(5), 1446-8.
92
van de Pol, G., van Brummen, H.J., Bruinse,H.W., Heintz, A.P., van der Vaart, C.H. (2006). Psychosocial factors and mode of delivery. Journal of Psychosomatic Obstetrics and Gynecology, 27(4), 231-236. Waldenstrom, U. (2004). A negative birth experience: prevalence and risk factors in a national sample. Birth, 31(1), 17-27. Ware, J.E., Jr. & Gandek, B. (1998). Overview of the SF-36 Health Survey and the International Quality of Life Assessment (IQOLA) Project. Journal of Clinical Epidemiology, 51(11), 903-12. Wisner, K.L., et al. (1999). Obsessions and compulsions in women with postpartum depression. Journal of Clinical Psychiatry, 60(3), 176-80. Wisner, K., Parry, B. & Piontek, C. (2002). Postpartum depression. New England Journal of Medicine, 347, 194-198. Wisner, KL, et al. (2004) Prevention of postpartum depression : a pilot randomized clinical trial. American Journal of Psychiatry, 161, 1290-1292. Wisner, K. L., Hanusa, B. H., Peindl, K. S., & Perel, J. M. (2004). Prevention of postpartum episodes in women with bipolar disorder. Biological Psychiatry, 56(8), 592- 596. Wisner, K. L., Hanusa, B., Perel, J.M., Peindle, K.S., Piontek, C.M., Sit, D.K., Findling, R.L., Moses-Kolko, E.L. (2006). Postpartum Depression: A randomized trial of sertraline versus nortriptyline. Journal of Clinical Psychopharmacology, 26(4), 353-360. Wu, J., Vigurea, A., Riley, L., Cohen, L., Ecker, J. (2002). Presented at the 22nd Annual Meeting of the Society for Maternal-Fetal medicine, New Orleans, LA, January 14-19, 2002. Zekowski, E.M., O'Hara, M.W., Wills, K.E. (1987). The effects of maternal mood on mother-infant interaction. Journal of Abnormal Child Psychology, 15(3), 361-378. Zimmerman, M. , Posternak, M. A., & Chelminski, I. (2005). Is the cutoff to define remission on the Hamilton Rating Scale for Depression too high? Journal of Nerve and Mental Disorders, 193(3), 170-175. .
93
